Latest / Found in Interpretation Podcast / Ep. 40 - The Future of Interpreting: Humans vs AI | Dr. Jonathan Downey

Transcript

Ep. 40 - The Future of Interpreting: Humans vs AI | Dr. Jonathan Downey · Found in Interpretation Podcast

• 0:00
  Hello, hello and welcome to another episode of found in
• 0:03
  interpretation podcast. Once again, I'm one of your Co
• 0:06
  hosts and I'm you're Brian. Brian, how you doing?
• 0:09
  I'm doing fantastic. And we also have with us
• 0:12
  Jonathan Downey for a second time.
• 0:15
  I feel like the last time we just hit the tip of the iceberg,
• 0:17
  so I'm looking forward to this one.
• 0:18
  This is gonna be awesome. The they let me come back.
• 0:21
  Yes. Yes, absolutely.
• 0:23
  So don't forget everyone like share, subscribe, comment, it's
• 0:28
  so important every, every comment and especially every
• 0:32
  time you clips you click subscribe.
• 0:36
  It's really important for us. That's how the algorithm will be
• 0:41
  able to tell you that another video is coming up.
• 0:45
  So let's kick it. Alright, alright, we're, we're
• 1:00
  laughing in the background. You know, we were while you guys
• 1:03
  were listening to that little intro music, Jonathan.
• 1:07
  Jonathan is the author of several books on interpretation.
• 1:14
  I've got this one right here with me, which is Interpreters
• 1:18
  versus machines. Can interpreters survive in a,
• 1:23
  in an AI dominated world? Very, very current.
• 1:28
  And it was written in 2019. And then came its sequel, Doctor
• 1:34
  Downey and Mr. Stick, which is this is my iPad and this is an
• 1:40
  ebook which is a lot more accessible, easy to find and a
• 1:45
  lot cheaper than the first one, if I'm not mistaken.
• 1:48
  Because basically this, the second one is sort of a sequel
• 1:53
  to the first one, right? It's a sequel on the follow up,
• 1:56
  uncovers things that weren't really available to be looked at
• 2:00
  properly with the first one. So with the first one, computer
• 2:03
  aided interpreting tools were beginning to start being
• 2:06
  developed and I had a very utopian view.
• 2:08
  I was imagining like having a Google Glass thing over my eyes
• 2:12
  that gave me terminology and stuff.
• 2:14
  And now we've got more research out and we've got more data
• 2:17
  available now we can actually think about them, hopefully.
• 2:20
  But the I was shocked at how the basic presets, a basic
• 2:24
  understanding of interpreting throughout both books is almost
• 2:27
  exactly the same. Interpreting hasn't changed
• 2:30
  basically a huge amount for thousands of years.
• 2:33
  Sure enough, yeah. And I think your book also came
• 2:36
  like before, like ChatGPT and everything 2019.
• 2:39
  So you're kind of ahead of your time.
• 2:43
  Only ahead of my time for book sales, yes.
• 2:45
  Yes, for book sales. But I think also at that point
• 2:48
  it was a theoretical issue. So there were companies who
• 2:52
  claimed that they could match us.
• 2:54
  And I think everyone knew in 2019 that claims to be as good
• 2:57
  as interpreting weren't that great and we're almost certainly
• 3:00
  wrong. More people are convinced today.
• 3:03
  And so I think I had to rewrite the book and go going to rewrite
• 3:07
  the not rewriting the book, but write a sequel saying here's
• 3:10
  where we are. And I think what was helpful for
• 3:13
  me was writing that from the perspective of what do we know
• 3:15
  about interpreting rather than what do we know about AI.
• 3:19
  Too many people are assuming that if you know enough about
• 3:21
  AI, you understand interpreting. Yeah, you've got the wrong way.
• 3:24
  Around this one, yeah. You start with the tariff thing
• 3:27
  and then you say, well, from what I know about interpreting,
• 3:30
  what can we say about AI being used in?
• 3:32
  Interpreting. So what do we know?
• 3:34
  What about interpreting that changes?
• 3:36
  Yeah. What we know about interpreting
• 3:38
  is that interpreting is about people before us, about
• 3:40
  language. So interpreters are not picking
• 3:44
  up meaning that the speaker drops, carrying it across like
• 3:47
  Santa's little elves and then dropping them on the other
• 3:49
  speaker. Absolutely not.
• 3:51
  We're not carrying anything over being bridges of any sort.
• 3:54
  We're working in partnership with the speakers and the
• 3:58
  audience to generate and discuss meaning together.
• 4:02
  Now, that sounds really theoretical and abstract, but
• 4:05
  really what it is, is we are part of the conversation and our
• 4:08
  decisions are affecting how people understand.
• 4:12
  And then we see, especially if you've done maybe QA sessions or
• 4:15
  you've interpreted groups that involve discussion, you can see
• 4:18
  how the way the interpreter acts to one thing changes how the
• 4:22
  conversation goes. We don't, we're not always aware
• 4:25
  of that and simultaneous interpreting, but the same
• 4:27
  effect has been proven is there, that our decisions affect the
• 4:31
  course of the meeting, which then change how people
• 4:32
  communicate. So we're not carrying over
• 4:36
  objective meaning, mostly because objective meaning
• 4:38
  doesn't exist. We're working with people to
• 4:41
  generate and create meaning together.
• 4:44
  My senior PhD supervisor, Graham Turner had this great line.
• 4:48
  He said interpreting is the negotiation of meaning.
• 4:53
  OHK OK. Then you realise that that makes
• 4:56
  interpreting a lot more complicated than it would be if
• 4:59
  we were just the source text. Speaker says something, we pick
• 5:02
  it up and carry it across a bridge to the other side.
• 5:04
  It's not what we're. Doing, which is kind of what
• 5:06
  you're dispelling in your in the book.
• 5:09
  And because there used to be like the theory of the triangle,
• 5:12
  right, that you that you mentioned where like we're kind
• 5:14
  of like the in between. There's like a personal
• 5:18
  language, a person language B. You're the guy in between.
• 5:21
  But you've kind of dispelled that and said like, no, no,
• 5:23
  we're more. You did like a more pyramidal
• 5:27
  figure and. Explain that wasn't at first,
• 5:29
  even before the triangle was in it, just a straight line which
• 5:33
  was A&B and the interpreter in the middle.
• 5:35
  Yes. So the first model was the
• 5:38
  conduit model. I checked which dates is
• 5:40
  probably ancient, but certainly the way that we understand the
• 5:43
  conduit model now dates back to Claude Shannon's model of
• 5:46
  communication in the 1950s, which underlies how computers
• 5:49
  work now. So we're going to the 1950s
• 5:52
  early computing and that model was very much the speaker says
• 5:56
  something, the interpreter understands what the speaker
• 5:58
  says and says exactly the same thing to the the target text
• 6:02
  here or here if sign. Language, so they're the conduit
• 6:05
  to communication. That is the conduit model, and
• 6:08
  interpreters were literally seen as literally pipes.
• 6:11
  Whenever you hear discussions about accuracy and interpreting
• 6:16
  or how good we are with terminology or even things like
• 6:20
  how good they're propositional accuracy is, the model
• 6:22
  underlying that is a conduit model.
• 6:25
  And the conduit model says any additions or omissions, changes
• 6:28
  of registered, changes of tone summary that's a fault and that
• 6:33
  was wrong. From 1989 we can date there's a
• 6:37
  very precisely because because I interviewed the people behind
• 6:40
  it. From 1989, we realised first in
• 6:43
  sign language interpreting, then strangely enough from court
• 6:46
  interpreting and then we we go further out.
• 6:49
  Conference interpreting comes in the mid 2000s.
• 6:51
  We realized that doesn't work for interpreter.
• 6:54
  It doesn't describe in research terms, it didn't describe
• 6:57
  adequately what we were seeing. And so the triad model came in
• 7:01
  the late 80s to mid 90s. They they see interpreting as a
• 7:04
  triad. So we're all working together,
• 7:06
  we're understanding together. And then recently I was reading
• 7:10
  the research and realizing that common way of seeing the triad
• 7:13
  as we concentrate on this is the triad.
• 7:16
  It's the interpreting situation. If you read there, we try our
• 7:20
  reset. That's not really fully what it
• 7:22
  was saying. They concentrated on the
• 7:24
  situation, but they were aware of things like differences in
• 7:27
  power differences and how well people can communicate.
• 7:30
  They were awaiting the context that was going on, The aim of
• 7:33
  the meeting cipher to help people get away from.
• 7:35
  We're just thinking about the interpreting situation to do a
• 7:39
  pyramid model where you say every decision is based on
• 7:41
  what's going on in the meeting, but also the context in the
• 7:44
  back. I love this term back story
• 7:46
  behind this. Every time you interpret, we're
• 7:49
  being inserted into a story that has been going on in some cases
• 7:53
  for years, and we land in the middle of that story or
• 7:56
  sometimes at the end of that story, and we have to interpret.
• 7:59
  And how much of that story do we know?
• 8:01
  We're 510%. But that story is what's driven
• 8:05
  the meeting, which is what we interpreting at.
• 8:07
  And everyone in the meeting comes with their own back story,
• 8:10
  their own context, their own intentions.
• 8:12
  And we make our decisions based on what we perceive of the
• 8:16
  interpreted situation and our own background, their own story,
• 8:19
  our own intentions. You see, interpreting is that
• 8:23
  complicated. Then you go.
• 8:24
  It cannot be anything but human. Hmm.
• 8:28
  But when it cut, there must be a reason why that very first
• 8:34
  linear model is so attractive still today, although it dates
• 8:39
  70 years. It's very easy to understand and
• 8:42
  it's based on what I think was in 1979 paper by Michael Ready.
• 8:46
  He calls it the fork model of communication.
• 8:49
  The English language has am, I guess we say things like am I
• 8:52
  get I couldn't get across to it, or I couldn't get my ideas
• 8:55
  across or we're used to talking about communication as this
• 9:01
  linear thing. All of our metaphors about
• 9:05
  communication are about, you know what, what, what did you
• 9:08
  say to him? All of our communication
• 9:10
  metaphors in English and they need to think, think about
• 9:13
  French, but communication metaphors are linear.
• 9:17
  So our brains and our coach are conditioned us to think about
• 9:20
  communication being linear. So you know, if you're there's a
• 9:23
  classic painting thing is a lot of parents when they have
• 9:26
  children think that the way to get your kids to do chores
• 9:29
  around the house is to just say tidy your bedroom the
• 9:33
  successfully on that is. Well, anyone who has teenagers
• 9:36
  will know that doesn't. Work.
• 9:38
  Yes. And what did we learn?
• 9:40
  That you more effectively teach children things by being with
• 9:45
  them and listening to them and being present and actually
• 9:47
  having a conversation. Seeing tidy your bedroom is the
• 9:51
  least effective method to get any child to tidy their room.
• 9:55
  Yeah, yeah, totally, totally. And so we we, so we our brains
• 9:58
  also know that linear communication is the least
• 10:00
  effective way to communicate. But we still do it.
• 10:03
  But we still, but we still do it and we still, and so it's only
• 10:06
  when you get people to think through, well, hold on, you
• 10:09
  know, that doesn't work. And it is things like I, I, I
• 10:12
  was in my wife was working for a boss and he was dictating a
• 10:17
  letter and he used the word for match.
• 10:21
  No, just Scottish. That's not a problem.
• 10:22
  That just means I'm a fighter. An argument.
• 10:25
  She'd never heard that before. Because she's.
• 10:27
  English scrimmage. Scrimmage, scrimmage believe are
• 10:31
  not Scots. English has a lot of words for
• 10:33
  arguments. I am not going to guess why.
• 10:37
  We also have a lot of words for rain and quite a few of being
• 10:41
  drunk. Please do not do the whole Sapir
• 10:46
  Wharf hypothesis and hypothesis and tell me how Scottish people
• 10:49
  think I'm teetotal and peaceful. So but, but, but we, it is true
• 10:56
  that we are shaped by the way that we learn to think about
• 10:59
  things. Yeah.
• 11:01
  And most especially most monolinguals think.
• 11:04
  And I was explaining to someone. So I was in a chat with a guy
• 11:07
  who's only speaks English and he said to interpreting must be
• 11:10
  easy. I said, hey, he said, well, you
• 11:13
  just hear the French word, then you say whatever the English
• 11:15
  word is for that French word. He said like the the the French
• 11:19
  for Windows finette. And I was like, depends on the
• 11:22
  kind of window. Could be quiche, could be, could
• 11:25
  be train, you know, I mean, there was two different ways
• 11:29
  that you can see window in French depending on the context.
• 11:32
  Yeah. And suddenly it's like, so you
• 11:35
  just, it's not just one word for word.
• 11:37
  That was like that would make my job a lot easier.
• 11:39
  He really thought he. Really thought I know.
• 11:42
  But but but but that he he doesn't even understand his own
• 11:45
  language. I mean, look at an English
• 11:47
  Dictionary and you have like 10 definitions for something.
• 11:50
  Yes, but most the the folk definition of language and the
• 11:54
  folk we've understanding translation interpreting is.
• 11:56
  I still have to explain to people why word for word
• 11:58
  translation isn't a thing. And so when you have these folk
• 12:04
  views of language, I think Michael Ready calls them, then
• 12:07
  we're not battling against people who it's not that we're
• 12:10
  talking to people who are being deliberately obtuse, is that
• 12:15
  their culture has programmed them to see translation
• 12:18
  interpreting in certain ways, especially interpreting.
• 12:21
  And so to to explain to people why machine interpreting doesn't
• 12:26
  replace us, you have to change how they view interpreting to
• 12:28
  change how the view interpreting sometimes takes the change in
• 12:31
  how they view language. Yeah, but how you're, you're
• 12:34
  putting, I'm very surprised about this because you seem to
• 12:36
  be putting a lot of weight on culture to explain the reason
• 12:41
  why there's certain perception about interpreting.
• 12:44
  But this is not common only to the English speaking world.
• 12:48
  If you if you go to China, I'm sure that they'll they might
• 12:53
  also have a very linear way of seeing interpretation if they
• 12:57
  don't know about it. So do you mean like culture?
• 13:03
  What do you mean by culture? You were like human culture
• 13:05
  because it's not really specific to Anglos action.
• 13:08
  Culture. Well, there was some research
• 13:10
  that the word, the kind of terminology used to explain
• 13:14
  translation interpreting does affect how people say, I can't
• 13:16
  remember which culture, but there's one culture where their
• 13:19
  word for translation is more to do with, um, doing flips or
• 13:23
  turning up, turning something upside down.
• 13:25
  Oh cool. So there are there, there has
• 13:29
  been research on the metaphors that people use.
• 13:31
  So certainly Anglo-Saxon culture has a very linear view and
• 13:34
  you'll probably find that different cultures have some
• 13:36
  kind of other view. And I think also this linear
• 13:39
  view of communication, there's a kind of, there's a kind of basic
• 13:43
  logic to it that sounds like it should be right.
• 13:46
  And so I think a lot, but I would not be surprised if
• 13:49
  anthropologists were to go to other cultures and find that
• 13:52
  actually cultures that associate.
• 13:54
  I mean, even some cultures does one word for translation and
• 13:57
  interpreting. That's going to create a
• 14:00
  different metaphorical map in your mind than in English, where
• 14:03
  there are two words but most people don't realize the other
• 14:05
  one exists. Right.
• 14:07
  And that this is the whole and also.
• 14:09
  Things like in French, I thought Brad could mean sometimes people
• 14:12
  think I'm a singer because, you know, it's.
• 14:16
  In like in you and I are gonna go, Alan, you and I are gonna go
• 14:21
  to a conference next month up north and Yellowknife and
• 14:23
  Northwest Territories. And apparently, like in some of
• 14:27
  the Inuit languages, there's like 70 or 80 different words to
• 14:31
  describe snow. So like, it's, this is exactly
• 14:35
  what Jonathan's talking about. Like, it's, which one do you
• 14:38
  pick? Right?
• 14:38
  Like, like, and it's all about the culture.
• 14:41
  It's all about the context, like.
• 14:43
  And also, even going away from the word level, how many
• 14:46
  different ways are there in most languages of asking someone how
• 14:49
  they are? And I was there was a linguist,
• 14:53
  I can't remember who it was. I was talking to a linguist once
• 14:55
  who said if it. I think it may be in Cynthia
• 14:56
  Roy, who said every culture that has been examined there, they
• 15:01
  have a routine way of asking someone how they are, which
• 15:04
  doesn't actually mean that they're trying to get
• 15:06
  information on the person's mental or physical state.
• 15:08
  It's just a greeting. Yeah.
• 15:10
  And so every language that they've looked at apparently has
• 15:13
  equivalent of high. How are you?
• 15:14
  I'm fine. Or in America.
• 15:16
  How high? How are you?
• 15:17
  I'm great. I'm a billionaire.
• 15:18
  It's the same thing, but this is the the issue is that when you
• 15:25
  realize that someone saying hi, how are you doesn't mean that
• 15:28
  they want you to stop them in the street and give give you an
• 15:30
  8 hour discussion as to how they actually are.
• 15:34
  But most people go away. Yeah, how are you means how are
• 15:37
  you? No, it doesn't mean how are you
• 15:38
  people. Eventually you get them to
• 15:40
  realize that. And a lot of comedians have made
• 15:42
  their stock and trade the difference between what we see
• 15:45
  and what we mean. Well, once people realise that's
• 15:47
  going on, it's easier to explain then how an interpreting what
• 15:50
  someone says and what they mean can be two very different
• 15:54
  things. And depending on the context, we
• 15:55
  might have to go with what they mean rather than what they say.
• 15:58
  Yeah, well, it's a lot different too.
• 16:02
  Like say, for example, the gun contacts, like if a lot of the
• 16:05
  the deals that get done, they get done in the hallway, right?
• 16:08
  They don't get done at the at the table.
• 16:10
  So there's that too, like the hallway conversations and the
• 16:13
  side deals and all that. Like it's all part of the part
• 16:17
  and parcel of our job is to understand what's happening
• 16:20
  behind the scenes and then bring it to the forefront, to the
• 16:24
  general meeting, I guess between.
• 16:26
  The two parties I was practicing interpreting a preacher who was
• 16:30
  going on about how glad he was that the Clemson Tigers won.
• 16:35
  Outside of US college football, I hope no one who would have a
• 16:39
  clue who that is. Yeah.
• 16:41
  Where? So as an interpreter, especially
• 16:44
  in church, I've got to work out how do I make that make sense
• 16:47
  for this audience? What negotiation do I need to do
• 16:50
  in the time allotted to me in the context that's here, to make
• 16:55
  that work? Hmm.
• 16:59
  So now that we've defined the, the what interpreting is, I
• 17:06
  mean, we've, we've mostly done that now because this is one of
• 17:09
  the, and this was the first part of your book where basically you
• 17:13
  were you, you were explaining the different methods and
• 17:16
  conduits and, and all of that. And basically, I think that if I
• 17:21
  can boil down into something, it was that since the situation, I
• 17:27
  mean, since AI interpretation follows an old model and 70
• 17:34
  years later we're way beyond that.
• 17:39
  That's why I interpreting cannot work.
• 17:43
  That's and so if I, if I had to be the devil's advocate here
• 17:50
  because I'm trying to put myself in the shoes of a, of a client
• 17:54
  or a, some kind of software company or whatever, they would
• 17:58
  say, well, what matters is the result.
• 18:02
  And the result is that you say something to the machine and the
• 18:06
  machine will produce a content that is understandable to the
• 18:10
  listener. Whatever model we can create
• 18:13
  amongst ourselves is. It is.
• 18:16
  It is. Doesn't it boiled down to that?
• 18:19
  So there's a couple of things going on. 1st, I think most
• 18:21
  clients the results they want is not is the interpreting good but
• 18:25
  that the meeting work. And so for instance, if I when I
• 18:29
  was interpreting at that business negotiation, I
• 18:30
  mentioned in the book, the outcome that the client wanted
• 18:33
  was that they could sign off on the contract.
• 18:36
  How interpreting got them there was almost immaterial to them.
• 18:40
  And I think the quicker that we realized that interpreting
• 18:42
  always exists for a reason, the quicker we get better interpret.
• 18:46
  The second thing is that we're now realizing the AI
• 18:49
  interpreting is not doing a reliable job of putting the
• 18:54
  content over and away that that people can understand.
• 18:57
  There was a recent World Health Organization study which I said
• 19:00
  which I think they said that the accuracy of the eye interpreting
• 19:05
  range from a thing. I think it was like 5 to 80%
• 19:08
  We've also. Seen that's that's that
• 19:11
  everything and nothing. Yes, and all of the 60 speeches
• 19:14
  that they tested it on, only one would have reached the pass rate
• 19:18
  for the their local university final and everything exam.
• 19:22
  OK. And we saw things like getting
• 19:24
  the United States and Hamas mixed up, getting people's
• 19:28
  genders mixed up, calling one of the the senior officials our
• 19:31
  beloved African with there are serious reputational risks to
• 19:37
  using AI interpreting. Some of them are simply because
• 19:41
  I believe are simply because of things like, you know, needing
• 19:43
  to improve speech recognition and speech synthesis.
• 19:46
  And the, the fact that with in a interpreting the call it the
• 19:50
  cascade model, we would call it the conduit model and the
• 19:52
  cascade model. An error in one stage creates an
• 19:55
  error in in in another stage to sort things compound.
• 19:59
  So of course that was going to be an issue even if you had
• 20:02
  perfect speech recognition, perfect machine translation,
• 20:06
  perfect speech synthesis. The problem for clients is
• 20:09
  you're still a reputational risk because AI does not know what to
• 20:13
  do in situations that could put you, that could cause you
• 20:17
  problems. It doesn't know what to do when
• 20:20
  someone says something that's accidentally offensive.
• 20:23
  I was talking to an AI company that that specializes in
• 20:28
  interpretation, simultaneous interpretation, and they were
• 20:32
  telling me that one of their competitors lost a human.
• 20:35
  Yeah, it was a human that I was talking to.
• 20:37
  I checked how many fingers he had on each hand.
• 20:40
  But but he he was saying that one of his competitors had lost
• 20:45
  the contract because it miss interpreted a word that was a
• 20:52
  very racist word and it is not what the CEO of the company
• 20:55
  said. And so like it was halfway
• 20:58
  through his speech and then he had to spend the other half of
• 21:01
  his speech like apologizing for something that the AI did, not
• 21:05
  something that he said. So like you like, instead of
• 21:09
  going like finishing his presentation, he had to spend
• 21:12
  most of the time just apologizing for what the I
• 21:15
  thought he said. When things like I promise you
• 21:17
  I'm not I would never say that word.
• 21:19
  I'm not racist. I'm not this, I'm not that.
• 21:21
  So like. If it happened in Canada, he had
• 21:23
  to apologize twice as much. That's right, yeah.
• 21:26
  If it happened in France, has the quote a dead philosopher as
• 21:28
  well? Yes, exactly.
• 21:31
  Preferably a dead white male philosopher.
• 21:33
  Yes, exactly. The the The thing is, is that a
• 21:36
  lot of clients are risk averse. So clients want results and they
• 21:39
  want low risk. No.
• 21:42
  The interesting thing with humans is that if you're a human
• 21:46
  interpreter and you're a professional, you have signed
• 21:48
  the code of conduct, or at least you should have.
• 21:51
  If you've got good association, you also to an extent we are
• 21:56
  well, we are accountable for what we see and if there's an
• 21:58
  issue, there's a whole association that the clients can
• 22:01
  go to and we have to own up to our mistakes.
• 22:06
  Whenever I've looked at the terms and conditions of AI
• 22:08
  interpreting apps or platforms and someone can correct me if
• 22:12
  their platform is different, I have seen somewhere in the T's
• 22:15
  and C's. We take no responsibility trying
• 22:19
  to reconnect and come back. Yep, something.
• 22:23
  OK, so interpret human interpreters own up to their own
• 22:29
  issues. You know, we have
• 22:30
  accountability. If we do something wrong,
• 22:32
  there's an association that client can go to and the
• 22:35
  association can do various things.
• 22:37
  Someone can correct me if I'm wrong, but every EI system that
• 22:40
  I have looked at and I've checked the T&C's, they have a
• 22:44
  condition somewhere that says that they take no liability for
• 22:48
  harm caused by their system if they are not prepared to own up
• 22:53
  to the accuracy of their own system and the reliability of
• 22:56
  their own system. Should anyone else.
• 23:00
  I mean our our contracts will have liability limiting clauses
• 23:03
  and they're there for good reason, but still we see if
• 23:06
  there's an issue. Here's who you go to if you have
• 23:10
  an issue with an AI interpreting system.
• 23:12
  All the ones I've read so far please come at me if I've made a
• 23:15
  mistake but everyone I've read so far has basically said tough.
• 23:19
  Yeah, your Sol. So hold on a minute, if you're
• 23:25
  going to be rescuers, which probably corporate and medical
• 23:28
  clients should be in legal clients should be the lowest
• 23:31
  risk solution is a qualified human.
• 23:34
  Yeah, in in every situation. The lowest risk situation
• 23:38
  solution isn't qualified. And even if we look at cost
• 23:41
  benefit analysis, so the cost of, so you can get machine
• 23:46
  interpreting for free and you get what you get and it could
• 23:49
  break down halfway through or you can get an enterprise
• 23:55
  package or you can get a specific AI for interpreting, AI
• 23:58
  interpreting solution and you can pay for that.
• 24:01
  Well, some of the ones that I've seen are going to give you like
• 24:05
  4050 languages whether you need them or not.
• 24:08
  And at what point did the court does it actually become cheaper
• 24:11
  than humans if you're having to pay for the I interpreting
• 24:14
  solution anyway and you're paying for it knowing that
• 24:17
  you're liable for if anything goes wrong?
• 24:19
  Yeah. The ends up being the humans,
• 24:22
  although we look expensive upfront, would actually cheaper
• 24:25
  because you're less likely with a qualified competent human to
• 24:30
  have to call the lawyers or the PR people afterwards to make up
• 24:33
  for the message just happen. Yeah, yeah.
• 24:36
  But I don't think that most people think like that.
• 24:37
  People think about those issues only when they arise and only
• 24:42
  once they screwed up. Even if you try to warn them
• 24:46
  beforehand. I think it's human nature just
• 24:49
  to try it anyway until something fails.
• 24:51
  Yes. And to be honest, we, we've all
• 24:54
  done that kind of thing. And, and I think we need to
• 24:57
  allow for people to do that. I don't think it's wrong for
• 24:59
  clients to say we need to try. I I think we shouldn't be
• 25:02
  demonizing clients because, you know, Microsoft Translator gave
• 25:06
  them a good deal. You know, how many people do you
• 25:08
  know who won't go for a good deal to get something cheaper?
• 25:12
  Of course I tried, I tried the Apple, Apple, well, now the
• 25:15
  latest OS and the beta version that I've downloaded on both of
• 25:20
  my computer and my iPhone and there's this live translation
• 25:23
  feature. They, that's how that's the term
• 25:25
  they used live translation. And it, it, it feels like, I
• 25:33
  don't know, it felt clunky. It felt unnatural.
• 25:37
  It it's just and I called it a toy.
• 25:43
  So I think. It's not professor, not not for
• 25:46
  professional use. We have to be aware though that
• 25:51
  there are places where AI is the right solution.
• 25:56
  There are a lot of places where interpretation just aren't
• 25:58
  available. And I was, I remember getting
• 26:01
  into conversation with someone and they are churches that are
• 26:03
  near to migrant resettlement centres.
• 26:06
  So these churches want to help the people in the community.
• 26:08
  They want to be a church that's open to anyone, but the language
• 26:11
  profile of their church can change from one week to another.
• 26:15
  I can in good conscience, as a professional, say you should be
• 26:19
  getting humans every week because they don't know what
• 26:21
  humans are going to get into. People walk into church.
• 26:23
  Yeah. Was it that the story you were
• 26:25
  saying about the, the, they they were bringing up bus with?
• 26:30
  I think it's in your book, right, that I read the story of
• 26:33
  there was a church that I wanted to bring some newcomers and some
• 26:37
  migrants to the church. And then basically there was, I
• 26:42
  would say, a cultural difference on how to book the church and
• 26:47
  about bringing some people in and they had to reserve in
• 26:50
  advance and all of that. And the area ended up being so
• 26:52
  complicated nobody showed. Up well, So what happened there
• 26:55
  was there was a misunderstanding.
• 26:57
  So different cultures have different conceptions of time.
• 27:01
  African time is a known thing and I say African time.
• 27:04
  That's a massive stereotype because Africa is a huge cunt.
• 27:08
  But the clash there was between the people who wanted to come to
• 27:12
  the church when they need someone to come to pick them up
• 27:14
  and the church that wanted everything to be reliable,
• 27:17
  knowing when people when people were going to be there.
• 27:20
  And I was actually the interpreter there and it was a
• 27:23
  very awkward conversation to interpret.
• 27:24
  And what went wrong. There wasn't an interpreting
• 27:27
  issue at all. It was purely A cultural issue
• 27:30
  of we're doing this and this is a chat I have, I have with
• 27:34
  people regularly. You can see we're we're doing
• 27:37
  this meeting in three languages and the provision of
• 27:40
  interpreting itself is really enough to make things work.
• 27:45
  So in churches, for example, if a church says we're going to put
• 27:49
  on interpreting for them and the interpreting is like a stepping
• 27:52
  stone towards people learning English and just being in the
• 27:54
  church like everyone else. That is different to we're going
• 27:58
  to have interpreting for the life of the church because this
• 28:00
  is, these are our values as as the church and we want people to
• 28:04
  flourish in whatever languages they're going to use.
• 28:08
  That's different. The second one works because
• 28:12
  then you say, well, the interpreting is not just there
• 28:14
  for us, it's there for them, it's there for us too.
• 28:17
  And then you get, well hold on, if we have people speaking 6
• 28:19
  languages in our church, maybe we can have the preaching on one
• 28:22
  of those other 6 languages and suddenly languages become a core
• 28:26
  part of the church. What happened in that church was
• 28:28
  the language outreach was never going to be a core part of the
• 28:32
  church because of various things.
• 28:35
  And so it was always going to be a cost benefit analysis.
• 28:38
  Is the hassle of doing this worth the benefit that we get
• 28:44
  with us? For those coming to the church,
• 28:45
  it was where can we find a place where we'll be accepted?
• 28:49
  And by the way, we are who we are.
• 28:51
  We were raised how we were raised.
• 28:53
  This is how we do stuff now. There's always going to be
• 28:57
  negotiation on both sides. That's just life, you know, You
• 29:00
  get married, there's negotiation on both sides.
• 29:02
  That's how relationships work. Yes.
• 29:05
  But because the if you like the the home leadership didn't want
• 29:13
  the addition of language to mean change and felt that the price
• 29:17
  of changing was unacceptable. It was always going to breakdown
• 29:23
  and I think this is where we need to realise that.
• 29:26
  I didn't put it in Doctor Danny, Mr. Sticks and I found the
• 29:29
  thing. I missed it out of interpreters.
• 29:30
  Firstly, machines. There are some clients for whom
• 29:33
  interpreting is just this thing we have to have, and
• 29:39
  traditionally those have been our bread and butter.
• 29:42
  You know, when we talk about interpreting in human rights, we
• 29:44
  talk about interpreting accessibility.
• 29:46
  We talk about interpreting and, um, legal.
• 29:50
  Requirements, equity and inclusion like yeah, you name
• 29:53
  it. Conclusion In Canada, because of
• 29:55
  official language act, there's some someone interprets so much
• 29:59
  interpreting, but because of Zoom, now it's, it's it's
• 30:03
  terrible because I don't know if some people are listening or
• 30:05
  not. And that, well, no, but I mean I
• 30:10
  don't want to digress too much, but I think it affects
• 30:13
  motivation. Oh yeah, so I wouldn't start a
• 30:16
  four day event at the Scottish Parliament.
• 30:18
  I think the interesting team was about 16 strong because they
• 30:21
  needed more than one team per language.
• 30:24
  And over the four days, don't think I saw anyone picking up a
• 30:26
  headset. Yeah, no, no, that's a whole
• 30:31
  other thing. And I'm not going down the
• 30:33
  symbolic interaction route. We did that on the previous
• 30:35
  podcast. Yeah, but This is why back story
• 30:38
  in context, determine everything and there will be some events.
• 30:43
  So like I've done some interpreting that looked on the
• 30:46
  face of it. So I don't know about Canada,
• 30:48
  but in the UK and Europe there was a thing for a while that all
• 30:52
  of the interpreters wanted to do the really prestigious events.
• 30:56
  So I remember doing an event, it didn't look that prestigious.
• 30:58
  It was an internal thing. Internal event has some church
• 31:02
  leaders and a place, Am I OK? Internal event, I can handle
• 31:06
  that. We're discussing migration
• 31:07
  policy. You look at technical internal
• 31:10
  discussion of migration policy. Doesn't it seem like the most
• 31:14
  prestigious event? I mean, you compare that to UN
• 31:16
  or U or whatever and I'm OK. Is is a church event?
• 31:19
  I like church stuff or will do it.
• 31:21
  That one event was more life changing for me and for the
• 31:24
  people who were there then most of the other interpreting that
• 31:29
  I've. Hmm.
• 31:31
  And that event was some of the most challenging and terrific
• 31:34
  thing that I've done emotionally, as much as anything
• 31:36
  else than some of the other events I've done.
• 31:39
  And sometimes the events on paper look like this is
• 31:42
  straightforward, can be really challenging and can we really
• 31:46
  rewarding. And the events that look, you
• 31:48
  know, the, the, the agency, I love interpreting agencies.
• 31:51
  I had an agency come through once and paid me like 50% extra
• 31:55
  because it was going to be challenging and it was the
• 31:56
  easiest Internet thing I've ever done.
• 31:59
  The the, the first rule of interpreting is when the client
• 32:01
  says it'll be straightforward, book your council and now you'll
• 32:05
  need them later. And when the client says this
• 32:07
  will be really tricky, you know it's going to be the best, it's
• 32:10
  going to be the easiest interpreting you've ever found
• 32:11
  this. That's just how it works.
• 32:13
  And when the when the speaker comes to the booth and says I
• 32:16
  will slow down. Sorry.
• 32:20
  I'm a really fast speaker but I'll try to slow down.
• 32:23
  OHS, it is like seeing a tutorial that says the easy way
• 32:26
  to use Microsoft Teams. Yeah, Teams for Dummies.
• 32:35
  I I I did a video a short on I think no, it was on my
• 32:39
  multilingual church channel. I did a short called Why
• 32:41
  Interpreters Don't Ask Creatures to Slow down because you know
• 32:46
  what? You'll get 5.
• 32:47
  You'll either get and I've interrupted front desk and
• 32:51
  speaker. The speaker has slows down so
• 32:53
  far, you hope that they remembered to breathe before
• 32:57
  they finished the sentence. You you could drink like 3
• 33:01
  glasses of water and they're still on the same sentence.
• 33:04
  Or you get the speaker who slows down for about 3 minutes,
• 33:07
  realises they're running out of time and so speeds up to twice
• 33:10
  the speed that they would have been before.
• 33:13
  It's just not good advice. I think a lot of advice that we
• 33:16
  give speakers is actually counterproductive.
• 33:18
  That's a whole other topic, but I think a lot of the ways that
• 33:21
  we advise speakers is parenting. Again, behavioral behavioralism
• 33:27
  doesn't work. When you say to someone, it's
• 33:30
  like the whole thing. If I say to you to to everyone
• 33:32
  listening, and to you too. Guys especially, don't imagine a
• 33:35
  pink elephant. Guarantee you've just imagined
• 33:38
  the big elephant seeing the speakers.
• 33:39
  Don't speak too quickly. I was thinking of the movie
• 33:43
  Dumbo immediately. Yeah, the pink.
• 33:46
  Elephant that's not pink. Behavior over the one he
• 33:49
  hallucinates. You worst way to try and change
• 33:53
  a thing is to try and change it on the level of behavior.
• 33:58
  Yeah. When you get underneath of the,
• 34:01
  why is this happening? What's going on You?
• 34:05
  I think it's. It takes so long.
• 34:08
  But also you can appeal to people the opposite way around.
• 34:11
  So rather than telling speakers if you do this it will be a
• 34:13
  mess, I now do funny videos of people seeing that at work.
• 34:18
  So I did one on why you shouldn't be just speech from
• 34:20
  manuscript. The best way to do that is to do
• 34:24
  a hilarious video of seeing wide showing by manuscript speed when
• 34:28
  people see it from themselves and they understand.
• 34:31
  I got an e-mail once I did a. I've got a short on my YouTube
• 34:33
  channel called You interrupt us, ask for notes.
• 34:39
  And it starts with this kind of preacher.
• 34:41
  I made myself a fake dog call this preacher saying if I don't
• 34:44
  think it was even a preacher, I was doing my business person.
• 34:46
  I said, you know, why do you interrupt us, ask for documents?
• 34:49
  Don't you already know all the words?
• 34:51
  And the interpreter says, yeah, we know lots of words.
• 34:53
  We know all the best words. And then I explain, you know, if
• 34:56
  you were to do a speech on three different topics back-to-back,
• 35:00
  would you have to prepare? Sure.
• 35:03
  So do we, and that 142nd video got past us to suddenly start
• 35:08
  sending the interpreters notes. That's why I love having you on
• 35:12
  our on our podcast, you give so many little tips and tidbits for
• 35:15
  interpreters. Don't tell people to do
• 35:17
  something. Show them why they should do it.
• 35:19
  Yeah. Preferably going back to your
• 35:21
  book, I thought that one of the the cool things about your book
• 35:25
  too, that is that I've got reading material for the next
• 35:28
  few years because of the amount of of people you site on this
• 35:31
  book like now I'm curious to go and see other research, other
• 35:36
  literature, other like points of view like you.
• 35:39
  You've really done a great job with, with highlighting like who
• 35:44
  your references are and everything.
• 35:45
  And yeah, I'm, I'm super stoked to, to learn more.
• 35:50
  Like, even for like, I mean, I've been doing this for 22
• 35:53
  years, like, and I still want to know more.
• 35:55
  So I'm. Very keen and I'm always very
• 35:59
  keen. I remember that video recently
• 36:01
  on why a certain view of translation were wrong.
• 36:04
  I'm not a translator, but I know enough to be dangerous.
• 36:06
  I was a professional translator for four years.
• 36:08
  When I get trying to get my interpreting career started and
• 36:12
  someone said, well, that's just your opinion.
• 36:13
  And I was like actually, and I could list like 10 sources with
• 36:19
  I've tried to be with Doctor Diane and Mr. Styx.
• 36:23
  I have tried tried as far as possible to stay close to the
• 36:29
  biggest research consensus that we have with computer aided
• 36:33
  interpreting. That's a bit more difficult
• 36:35
  because I have my own concerns about how that software is being
• 36:38
  tested. For example, I I think the
• 36:42
  problem is, is that because this conduit model is so prevalent
• 36:47
  and because in terms of research methodology, testing and doing
• 36:53
  research under the triadic or pyramidal models is so
• 36:56
  challenging. I know researchers are doing
• 37:00
  really well with simulated interrupting to test.
• 37:06
  Computing these interesting under some of the conditions
• 37:09
  that are being tested and simulations is very hard.
• 37:11
  I've only just heard a few weeks ago it being used in a dummy
• 37:15
  booth, which are genius. We can still ask questions about
• 37:20
  motivation than we probably should, but for it to be used at
• 37:23
  a dummy booth and international organization is probably the
• 37:26
  closest we're gonna get to perfect results.
• 37:30
  Yeah, because you don't want, you know, doing it in live
• 37:32
  interpreting is ethically questionable.
• 37:38
  Again, you you I I could find ways of justifying it.
• 37:41
  You know I creative. Research.
• 37:42
  What do you mean by a dummy? Booth.
• 37:44
  So a dummy booth is a usually used to be a training.
• 37:48
  I think it's still used that way you have a real confidence, a
• 37:50
  real event going on, but you have an extra booth which has
• 37:54
  the same sound and video and feed as it looks functional just
• 37:58
  like a normal booth. But the microphone in the booth
• 38:01
  isn't connected to anything, so the interface doesn't interpret
• 38:05
  as normal. You can even, if you like, make
• 38:07
  the microphone light, light up as normal, but the sound techs
• 38:10
  have made sure the sound coming out of that booth is going
• 38:13
  nowhere. So it's often used as a training
• 38:17
  aid so that my university I train that will take people to
• 38:21
  dummy booth at the Scottish Parliament.
• 38:22
  They will be in a real interpreting booth with a real
• 38:25
  console and a real live meeting interpreting, but no one can
• 38:29
  hear them. Words, but it gets recorded so
• 38:32
  they can hear it after. It could be you can set it up to
• 38:36
  record if you want. You can set it up just as an
• 38:38
  experience if you want. With that study that I was
• 38:40
  talking about, they recorded it and they also had multiple video
• 38:43
  feeds so they could see how they interpreters were using the
• 38:46
  tool. In fact, I think, I believe I'd
• 38:48
  need to check, but I think there's, yes, they got the
• 38:50
  interpreters to where what are called eye tracking glasses as
• 38:53
  well. So they would get a live view of
• 38:56
  where the interpreters are looking.
• 38:57
  And they, they gave us a, a schematic of here's the booth
• 39:02
  wall. Here's where the interpreters
• 39:03
  were looking in the booth most often.
• 39:05
  They could tell when the interpreters were using the tool
• 39:07
  and when they weren't. When the interpreters were
• 39:09
  scanning the audience how they could, even from the eye
• 39:11
  tracking. I think they they may have done
• 39:14
  brain scans as well. They could tell you what the
• 39:16
  interpreters are using the tool for and how that's affecting
• 39:20
  their interpreting performance downstream.
• 39:23
  They look at whether they were looking at their phones.
• 39:27
  Yes, if they wish they could tell if they were doing Candy
• 39:31
  Crush in between chefs, yes Candy Crush.
• 39:34
  But so you that level of data you could theoretically do it in
• 39:39
  live interpreting. Hmm.
• 39:41
  The client may have something to say about that, OK.
• 39:45
  Because again, for clients and I think we've said this before,
• 39:48
  the main thing for the client is, is this going to do for me
• 39:51
  what I needed to do? You also have issues where if I
• 39:55
  researchers looking at the data, there are going to be
• 39:58
  confidentiality issues. Because the researcher has to
• 40:02
  record the speaker and the interpreters possible to do the
• 40:07
  institutional context. If you're doing, I don't know,
• 40:09
  business negotiation or you're doing, um, some sort of personal
• 40:13
  thing, you're not getting permission.
• 40:15
  You're just not. It's not happening.
• 40:18
  And even to get permission to record court interrupting for
• 40:21
  research, I've heard is is really, really difficult to do.
• 40:24
  And then you have rules as to what can you, what can you
• 40:28
  release to the public and what can't you researchers on?
• 40:30
  They're under pressure to to have what's called open data,
• 40:33
  where all the data they generate goes online.
• 40:36
  Again, I know most of my clients would just say it out right now
• 40:41
  for good reason. And so it's the, it's the
• 40:45
  nearest we're going to get to asking, is this stuff useful in
• 40:49
  real life? And the answer was, um,
• 40:54
  experienced interpreters using the two well, can actually
• 40:57
  become more accurate and can get names and numbers better, and
• 41:02
  seem to be using it in such a way as to reduce their cognitive
• 41:05
  loads or how hard they're thinking.
• 41:06
  That's great, but we know from data elsewhere that interpreters
• 41:10
  using computer aided interpreting tools tend to have
• 41:12
  worse intonation, and there are suggestions also that the
• 41:16
  strategies that they use in interpreting change.
• 41:19
  So if you're faced with a really long number, your reflex is a
• 41:23
  professional interpreter is. How much of this do I need?
• 41:27
  And an interpreter with our tool will naturally concentrate on
• 41:31
  the meaning of the number in the context.
• 41:32
  And so based on that will give the right level of, um,
• 41:37
  precision, if you like. So we know sometimes 10 million
• 41:40
  is 10 million, and sometimes people really need 9,999,999.
• 41:45
  We know reflexively that's how you do it as an interpreter.
• 41:49
  When I screen gives you the number exactly.
• 41:53
  There may be in the. Some suggestive results that
• 41:56
  interrupt us will go. Well, I've got the full thing in
• 41:58
  front of me. I might as well use which makes.
• 42:01
  Sense does the research can tell the did did they make the
• 42:05
  difference between something popping up because there is an
• 42:09
  aide versus let's say there's a PowerPoint presentation that is
• 42:14
  being displayed on the screen? So they were only looking at the
• 42:17
  effect of computer aided interpreting tools I think.
• 42:20
  That same number, you know what I'm talking?
• 42:22
  You know what I mean by that, right?
• 42:23
  Like. So, so that there's some
• 42:26
  research on how we're using AIDS and it seems one of the problems
• 42:30
  is what they call emphasize. So if they do research on how 2
• 42:34
  interpreters use the two on an event, it's good research.
• 42:37
  How representative is that of every other interpreter?
• 42:40
  Well, if they do 50 interpreters, for it to be
• 42:42
  manageable, they have to get them in the lab or heard
• 42:45
  representative, as the lab was really terrific.
• 42:48
  Yeah, there's no easy way out of that thing of if you want to be
• 42:53
  really realistic, you're gonna have a small number of
• 42:55
  interpreters. If you want to be really
• 42:57
  generalizable and to be able to see this is true of most
• 43:00
  interpreters, you're going to be moving away from real life.
• 43:03
  That's the nature of research. I don't fault people for that.
• 43:07
  I think you have to make different decisions depending.
• 43:10
  On. But there's other questions that
• 43:11
  could come in because when it comes to the reliable
• 43:14
  reliability of the sample, because I know personally that
• 43:18
  if I'm being asked to do to, to, to have a performance, basically
• 43:25
  it's a blank shot. Not going to perform the same as
• 43:29
  if I'm in a conference and I see 100 people there and I see
• 43:33
  people picking up some headsets. And then suddenly there's the
• 43:36
  adrenaline coming in and the excitement and suddenly I will
• 43:41
  want to perform a lot more than actually, I've noticed that I
• 43:44
  perform a lot more if I'm live somewhere or I'm more nervous
• 43:48
  maybe. But it's very different than
• 43:51
  being at home like in a in a safe environment, rather than
• 43:58
  being a like a thrown in the. What about?
• 44:03
  Right in the pit. What about picturing everybody
• 44:05
  naked? Does that help at all or not?
• 44:08
  There's two things going on there.
• 44:10
  One, I was really glad that I met up with some interpreters
• 44:12
  and I said I, I whispered in the rear.
• 44:14
  I said I'm you know, when you do practice interpreting at home,
• 44:17
  in half of them left because a lot of interpreters don't
• 44:19
  practice. I said, you know, when you're
• 44:20
  practicing interpreting, is it a lot worse than when you're
• 44:24
  actually interacting in real life?
• 44:26
  And they were like, yeah, I was like, kids, it's not just me.
• 44:29
  My, my interrupting practices are really not good.
• 44:33
  So they're much worse than I am in real life, again, for the
• 44:36
  back story reasons. And that the motivation that you
• 44:39
  just said, I think back story motivation explain a lot more
• 44:42
  about interpreting than we're probably ready to accept.
• 44:45
  And then the other thing is that I am now convinced there is no
• 44:49
  data for this. It's just my thing.
• 44:51
  I am convinced that remote simultaneous and then personal
• 44:54
  simultaneous are two different jobs.
• 44:58
  Remote simultaneous for me is at least partially managing the
• 45:02
  crazy streams of information that you have going on.
• 45:05
  When I'm doing remote I have a 2 screen setup.
• 45:08
  If you put eye tracking glasses on me, I would probably make
• 45:11
  them explode. The amount of movement that my
• 45:12
  eyes do, like whatever their measurement limit is, I'd be
• 45:16
  beneath that. Um, and also things like if
• 45:19
  you've got, there's a difference, I've noticed the
• 45:21
  difference between if I've got a church where it's just one
• 45:23
  speaker and one screen versus a Zoom call where you have 24
• 45:27
  people on the screen and any of them could speak at any moment.
• 45:31
  That to me is a difference. But then you put me in real
• 45:33
  life. The last time I was in real life
• 45:35
  was in one of those little half boot desktop boots things.
• 45:38
  Right, the table tops. Yeah, the table tops, it was so
• 45:40
  different being in a tabletop, being able to like elbow my my
• 45:43
  booth mate for a prompt or even like being able to talk to
• 45:48
  people at lunchtime and go, you know, how's it going, getting
• 45:51
  live feedback from the client. Also, the other thing is if
• 45:54
  you're in a booth, you can turn your head and you can see who's
• 45:58
  got a headset on, but you can also feel the atmosphere in the
• 46:02
  room change if you don't zoom call.
• 46:06
  It seems to me that I could be wrong, but I it's very difficult
• 46:09
  to read the atmosphere on a Zoom call and I've noticed that I get
• 46:13
  blindsided a lot more on a Zoom call than I do in real life.
• 46:15
  In real life, I can see something complicated or
• 46:19
  difficult coming a mile away because I've learned to read the
• 46:22
  room and I can see people. Shoulders move and I can see
• 46:25
  positions and chairs move, and I'm aware, and my interpreting
• 46:29
  instincts go. Something tricky is about to
• 46:32
  come get ready. The amount of times in Zoom
• 46:35
  where someone has pressed the button to speak, I don't know
• 46:38
  what they're going to say and they drop some big thing.
• 46:43
  If they remember to turn the mute everything off.
• 46:47
  Clients in person remembering that microphone exists, that's a
• 46:49
  whole other thing. My microphone tutorial is coming
• 46:53
  soon. No, but this thing of the unmute
• 46:55
  themselves and they say something and you have no
• 46:58
  warning as an interpreter what kind of things that person says
• 47:03
  and whether what they've said is just going to wreck the entire
• 47:05
  meeting. If you don't know, you can't
• 47:07
  wait. As if you're in the room and you
• 47:10
  see someone grab the mic and they do the leaning forward in
• 47:12
  their seat. Yeah.
• 47:14
  Or that thing or the other, that one you know, you know that
• 47:21
  maybe you should switch over with your birthdate.
• 47:24
  Yeah. You know there is a good time to
• 47:28
  change live interpreter where you go and find a coffee in a
• 47:32
  crash helmet. Yeah, I I wanted to one of the
• 47:37
  topics that I wanted to talk about when I read your book is
• 47:44
  peoples misconception about how the technology works beneath
• 47:53
  because you wrote your book, you wrote your first book in 2019.
• 47:59
  And I think that a lot of people would say 02019 we're in 2025.
• 48:05
  That's more than a lifetime ago. It's been 6 years.
• 48:09
  Technology must have changed so much and, and basically what
• 48:14
  you're saying is that yes, it's been 6 years but the technology
• 48:20
  may seem different but it's still the same underneath.
• 48:24
  Am I wrong? It's fundamentally this.
• 48:26
  So far, this is what I have to differentiate machine
• 48:29
  interpreting, machine translation.
• 48:32
  Machine interrupting or or speech translation as it should
• 48:35
  probably be called, is fundamentally working to the
• 48:38
  same basic cascade conduit model as it has been for as long as
• 48:43
  the technology has existed. There is ongoing work on what
• 48:47
  they call end to end models, but the last I heard, and again
• 48:51
  someone can correct me, the last I heard that wasn't ready for
• 48:55
  Showtime yet because the N 10 model is you take the machine
• 48:59
  translation thing out the picture and you go straight from
• 49:01
  waveform to waveform. If you've edited the podcast,
• 49:04
  then you've just done the sound. You know what a waveform looks
• 49:07
  like. And the idea is that you can get
• 49:08
  the waveform in French and the waveform in English and just
• 49:10
  match mathematically. It makes perfect sense and there
• 49:15
  are some advantages. If that works, there are some
• 49:17
  things that are going to get a lot better, like intonation,
• 49:20
  like emotion, representation of emotion should get better.
• 49:25
  But most and I believe all of the currently commercially
• 49:28
  available software uses the cascade model.
• 49:32
  And so no matter how good you get the machine translation,
• 49:36
  machine translation has moved on quite a lot in six years.
• 49:40
  And there we're working now is almost entirely, I believe
• 49:44
  neural or elims or probably some combination of both.
• 49:48
  So machine translation has moved on a lot.
• 49:51
  Machine translation quality estimation I believe has
• 49:53
  advanced. In fact, I found out it's
• 49:55
  advanced quite a lot. Interpreting is still stuck in
• 49:58
  this cascade model wasteland with all the same weaknesses.
• 50:02
  Eyes on 20/19. The bigger problem for me is
• 50:07
  that our ability to talk intelligently about what the
• 50:11
  software can and can't do, and our ability to be persuasive and
• 50:15
  entertaining is still the same as it was in 1955.
• 50:18
  There remained 2019 and every director who I believe has been
• 50:23
  on the show makes this point continually that the gap between
• 50:26
  what we do and what we see we do is as wide as it has ever been.
• 50:32
  And we now know pretty much for sure the what clients need to
• 50:36
  see and need to hear us do realistic account of what we
• 50:40
  actually do. You know, the stuff we share
• 50:42
  during conferences, the stuff we share if we we go to cafe was on
• 50:47
  and we've got a nice hot chocolate, nice sandwich and
• 50:50
  we're chilled out. Those stories, the ones we
• 50:53
  usually only share between interpreters are way like 10
• 50:58
  times more convincing than we are accurate.
• 51:02
  We're objective, we're impartial, and we now know that
• 51:05
  for sure. Now I would say like one of the,
• 51:08
  the, the things with AI that you pointed out and other guests on
• 51:12
  our podcasts have pointed out as well is in today's world of AI,
• 51:18
  like it's so important to always think that you're being listened
• 51:21
  to. And, and like, because like you
• 51:24
  said, like the practice is always worse than than live,
• 51:27
  right? So like I would say that like,
• 51:30
  no matter if you're on Zoom, there's only three people and
• 51:32
  maybe someone isn't listening or like you're at a conference, no
• 51:35
  one picks up headsets. You never know if somebody's
• 51:37
  going to be listening to you. So I think you should always
• 51:41
  keep in mind that you should always deliver your best
• 51:43
  performance. And I think Alan mentioned this
• 51:48
  before, like when he was doing his exit exam at Glendon
• 51:50
  College, he had a picture of his son and he was interpreting as
• 51:56
  if he was telling the story to his son.
• 51:58
  And I think that's so important too, is to have that emotional
• 52:01
  side of things where if you picture like someone you love,
• 52:05
  like your, your, your, whether it's a parent or spouse or child
• 52:10
  or partner, whoever it is, like just picture them and that
• 52:14
  you're telling the story to these people and, and bring that
• 52:17
  context out. Bring that emotion out, Bring
• 52:20
  that that extra thing that AI can't offer.
• 52:25
  So I think that's how we're going to maybe stand out from AI
• 52:30
  and that's the way that we're gonna convince people that this
• 52:34
  is the choice that people need to make in order to be
• 52:36
  understood properly. It's not just the conduit.
• 52:39
  It is more than that. It's your pyramid structure.
• 52:43
  It's the context. It's.
• 52:45
  All of these different things that are still one of my
• 52:49
  favorite lines, which I don't say enough, is that the the
• 52:51
  biggest advantage we have over AI is AI cannot cry and there
• 52:58
  are times as an interpreter even sometimes in the most technical.
• 53:02
  Nor can Mr. Spock. Spoke he gets the point far and
• 53:06
  said, but you know, if when people are trying to persuade.
• 53:11
  So I don't want to go into text types and genres and meeting
• 53:14
  texts, but we all know as interpreters, there are meetings
• 53:18
  that exist to give people information.
• 53:21
  And in many cases the meetings that just exist to give people
• 53:24
  information. At some point in time, AI is
• 53:27
  going to be delivering a lot of those.
• 53:30
  Because if it's just you need to know this, then fine.
• 53:34
  But even in information heavy meetings, I think I tell the, I
• 53:38
  kind of do the joke story in the Doctor Downing Mr. Stick is when
• 53:40
  an accounting gives up to stands up, not gives up.
• 53:43
  Maybe when an accountant stands up to give the financial state
• 53:47
  of our company, it's a rookie interpreter who says all that's
• 53:52
  being transmitted here is the financials.
• 53:55
  Yeah, the numbers. The the accountant is not just
• 53:58
  giving numbers. I mean, otherwise, it'd be
• 54:00
  present in the lottery, the account that's giving numbers
• 54:04
  which say something about the financial state of the company,
• 54:07
  which reassures people that the company is being well run and
• 54:10
  the accountants doing a good job.
• 54:12
  If the accountant walks into that meeting with their shirt
• 54:15
  saying, I love embezzlement, at least some of those messages are
• 54:18
  not going to arrive. Yeah, But yeah, but we as
• 54:22
  interpreters have to listen to the accountant and go, what does
• 54:25
  $6 billion mean in this context? What in what?
• 54:30
  Which of these figures do I need to give exactly Which of these
• 54:33
  do I need to round? Which of these have emotional
• 54:36
  resonance? And we are making those
• 54:39
  decisions all the time. And it's why I love interpreters
• 54:42
  specializing in finance, because I don't always pick up those
• 54:44
  clues. Clues in finance talks.
• 54:47
  And you're not doing justice to a talk if you're missing half
• 54:51
  the clues of what the person trying to communicate.
• 54:53
  So I have a a a booth mate who I used to work with regularly, who
• 54:56
  was a finance specialist and who could do these talks with the
• 54:59
  plum. And everyone's like other
• 55:00
  companies being wrong. Great.
• 55:01
  I'm like, well, how much of that is because my booth me Just make
• 55:04
  your accountant sound good, whereas if you there are certain
• 55:10
  speakers that if you give me the money, even if you get like
• 55:12
  emotionally charged topics, I'm used to handling that world.
• 55:16
  You know, you give me a preacher or you give someone has got
• 55:18
  something emotionally charged. I think I said last one, Artie
• 55:21
  or creative or even someone who demonstrating something.
• 55:24
  I've got it straight away because I can see as a public
• 55:28
  speaker myself, I can log into. I know what you're trying to do,
• 55:32
  right. I know how you do that in
• 55:33
  English. I know how you do that in
• 55:35
  French. I'm going to make sure I'm doing
• 55:36
  the same thing as you are, and it will mean that the posies
• 55:39
  land in different places. It will mean that sometimes I
• 55:41
  summarize, sometimes I have to explain some.
• 55:45
  We all know that. What it takes to get
• 55:47
  interpreting out the door. Yeah.
• 55:49
  And for clients to be happy, well, if we know that perhaps we
• 55:56
  need to help the world see that because the stuff that we were
• 55:59
  ashamed of is our only We have shown the world that AI isn't
• 56:05
  doing our job. The the the interpreter and I
• 56:09
  think I said the interpreters pressing machines.
• 56:11
  The interpreter who still says don't add or omit is about to be
• 56:14
  omitted from interpreter. They want their name will not be
• 56:19
  added anymore. Because the best don't add or
• 56:22
  omit is the I Yeah, if it's good and if this sounds.
• 56:26
  Good. Speaking of of that, are we, we
• 56:30
  haven't discussed this or anything, but are we should we
• 56:33
  give away the future of AI that? You right, that was the last
• 56:38
  part of the book. Yeah.
• 56:40
  Which one? Well, well, you did.
• 56:43
  I don't know if we're kind. Of like, I don't know if we're
• 56:45
  gonna give the punch line during the podcast or people should
• 56:48
  read it. It's up to you, Jonathan.
• 56:49
  I don't wanna give the fruit punch line away, but one thing
• 56:52
  that I do want to say is, and this is in the context of as
• 56:59
  we're recording this, about four days ago, there was a big furry
• 57:02
  on LinkedIn over a study released by Microsoft saying
• 57:04
  that they would take everyone's jobs.
• 57:06
  Yeah, there will be. By the time this goes live there
• 57:09
  will be a video on why that's absolute nonsense, but.
• 57:14
  OK, so Microsoft released a list where they're saying which jobs
• 57:19
  are going to be wiped out by artificial intelligence.
• 57:24
  Not change, not modify, not accompanied, wiped out.
• 57:29
  Saying that someone is talking nonsense is that they're making
• 57:33
  100% confident predictions about the future.
• 57:38
  Anyone who says this is what's going to happen in the future,
• 57:41
  it's like, mate, you don't even know what you're having for
• 57:44
  breakfast tomorrow. Give over like you don't know if
• 57:47
  you're gonna go to the cupboard tomorrow and you'll be at Coco
• 57:49
  Pops. Don't, don't go predicting the
• 57:51
  future on me. And like people don't, you don't
• 57:54
  even know the next time you cross the road whether you make
• 57:56
  it to the other side for sure. So don't go telling me that you
• 58:00
  know what's gonna happen in five years time when you have you
• 58:04
  don't. And so I think we shouldn't be
• 58:07
  confidently predicting the future.
• 58:09
  We should be saying that there are things that we should look
• 58:12
  at, that we should examine, that we that there are trends that we
• 58:15
  need to be aware of. Absolutely.
• 58:17
  I think it's good to say to people here are trends that we
• 58:19
  need to be aware of. I think also we need to be
• 58:23
  careful of learned powerlessness.
• 58:26
  If you say to people in five years there won't be any human
• 58:30
  interpreters, you're in danger of having a self fulfilling
• 58:34
  prophecy because people get discouraged and they leave.
• 58:38
  I don't see anything helpful for anyone to see interesting going
• 58:43
  to disappear in five years. I don't think that helps anyone.
• 58:46
  That doesn't help clients, doesn't help us, doesn't even
• 58:48
  help the tech companies. But hold on, maybe I'm missing
• 58:52
  something, but do interpreters actually say that?
• 58:56
  I have come across interpreters, so maybe not five years, who say
• 58:58
  they don't think interpreting has a future.
• 59:02
  I've also come across interpreters who are very
• 59:04
  pessimistic and have been saying that for 30 years, but they're
• 59:07
  still years still, yeah, regardless of whether AI is
• 59:12
  around or. Anyone in mind, Brian?
• 59:14
  Anyone. No names?
• 59:15
  No. Names I I see no upside to that.
• 59:19
  I see upside for preparing. I see upside of being aware,
• 59:23
  confidently predicting your own demise.
• 59:26
  Like what? On the other hand, I would say
• 59:30
  that we do need to be aware that some of our current predicament
• 59:34
  is our own fault. And there's people forget.
• 59:38
  People forget the the the difference between availability
• 59:42
  of technology and adoption. Yes.
• 59:45
  I mean, so maybe and I and we were talking about this and then
• 59:51
  in another episode with another guest, but basically self
• 59:57
  driving cars, maybe the technology is 90% there, but the
• 1:00:01
  last 10% will take as much time as the as the other 90%.
• 1:00:04
  And there's also regulations and ethical issues and insurance and
• 1:00:10
  all of that. So even if the technology would
• 1:00:14
  be available even 2030 years from now, perhaps it won't be
• 1:00:18
  something that is widely used, or maybe it'll be assisted
• 1:00:22
  driving just because people either aren't ready, they won't
• 1:00:25
  trust the machine, or. There are all kinds of rules and
• 1:00:28
  regulations and it's nice here in Quebec.
• 1:00:31
  So I think it's a bit of the same thing with not only
• 1:00:36
  interpretation, but also every technology.
• 1:00:38
  I mean, in interpretation we're pointing out so many things that
• 1:00:43
  don't come up to the mind of some agencies, client, etcetera.
• 1:00:48
  But once you start really putting your hand into it, then
• 1:00:53
  you you realize that it's not as easy as it sounds.
• 1:00:56
  When you don't know anything in life, it's easy to just brush a
• 1:01:00
  stroke and say, you know what? This is what's going to happen.
• 1:01:03
  But you know, these people are at Microsoft is basically doing
• 1:01:09
  they're they're just trying to raise money.
• 1:01:10
  It's mean. The reason it's PR, the reason
• 1:01:13
  why they're saying that, it's because people wanna say ohhhhh,
• 1:01:16
  see, Microsoft said it, so we should invest more money and
• 1:01:20
  their stocks will go up. And they are on AI.
• 1:01:23
  The revenue versus profit ratio is terrible right now.
• 1:01:28
  I, I read an article today that every single AI, every single
• 1:01:31
  large AI company is making a loss.
• 1:01:34
  And if it wasn't for continued investment, they wouldn't exist.
• 1:01:37
  So the business model relies on PR to get investment to keep
• 1:01:40
  them going. But didn't we say that for a
• 1:01:42
  social media or even for Google? Or so I believe.
• 1:01:47
  Google managed to turn a profit fairly early.
• 1:01:49
  Social media is is a different a different animal.
• 1:01:52
  But I think, you know, when it comes to technology and
• 1:01:55
  adoption, everyone says where's my flying car?
• 1:01:57
  The exist but the right. But you need that drivers
• 1:02:00
  license and the pilot's license for them.
• 1:02:02
  And you're not gonna like what you have to do about tyre
• 1:02:04
  pressures because aircraft needs much higher tyre pressures than
• 1:02:08
  cars. So you're going to have to soon
• 1:02:10
  as you land your flying car, you better get yourself down to a
• 1:02:13
  tire pump and take 3/4 of the way your tires owned.
• 1:02:15
  By the way, you need to swap the nitrogen in the aircraft tyre
• 1:02:18
  for air in the car tyre. So you're going to be illegal at
• 1:02:21
  some point, the same as we have nuclear power trains for a
• 1:02:24
  while. Whether they go just because I
• 1:02:28
  take note this techno futurism that says as soon as a
• 1:02:30
  technology exists, everyone adopts it.
• 1:02:32
  It's like, have you been very since the 90s?
• 1:02:37
  Like, really. Like, yeah, some things people
• 1:02:41
  realize that a really, really bad idea.
• 1:02:44
  Like someone made a phone with a with two screens and one of them
• 1:02:47
  swiveled round. I have kids.
• 1:02:49
  I give that 10 minutes. For the yeah.
• 1:02:53
  Before the break off. Well, it's like the CD man right
• 1:02:57
  after the the the Walkman. Like they just, If you want some
• 1:03:01
  music that skips while you're running, this is your adoption.
• 1:03:06
  I mean, we're playing myself. I don't, I, I, I bought portable
• 1:03:10
  CD player cause it was the cheapest way to play CD's.
• 1:03:13
  Yeah, I never made the portable. And so the technology can exist
• 1:03:17
  in people can go, yeah, that's nonsense.
• 1:03:18
  Or you end up with technology that becomes very niche.
• 1:03:21
  Apparently we all have 3D printers and we're printing out
• 1:03:24
  everything 3D at home. Don't even have a printer right
• 1:03:28
  now. We also.
• 1:03:30
  Have the technology. We have the technology from the,
• 1:03:35
  I was going to say we have the technology for reliable inkjet
• 1:03:37
  printers and find me one of those printer manufacturers
• 1:03:42
  could make a reliable printer, but there's no fine.
• 1:03:45
  So the this is The thing is that we need to separate the
• 1:03:49
  predictions for the sake of PR from the Deezer trends that are
• 1:03:54
  going on. And also, I think I said it in
• 1:03:56
  the last episode, I'm going to say it again.
• 1:03:58
  If we want to have a future, we should learn from blacksmiths,
• 1:04:04
  engineers, mathematicians, historians and learn how to talk
• 1:04:09
  interestingly about our work. I I was aware conference a few
• 1:04:14
  weeks ago and there was a, an ancient historian called Medi
• 1:04:17
  Beard and she was onto TV doing a program about ancient Rome and
• 1:04:23
  it got watched. There's a guy called Alex Steele
• 1:04:26
  who's a blacksmith. The guys like and he's 20s
• 1:04:29
  thirties, several million YouTube subscribers watching and
• 1:04:33
  make stuff in blacksmithing. Now, if you asked anyone in the
• 1:04:36
  street, the blacksmith still existed.
• 1:04:37
  So you know, they're all gone. Well, Alex still exists and is
• 1:04:41
  making quite a nice life for. Himself and feel like.
• 1:04:46
  We just need. Yeah, but you know, it shocks me
• 1:04:51
  how interpreters were communication experts.
• 1:04:54
  We make speakers more interesting for a living, even
• 1:04:57
  we're not supposed to sometimes was close over that, but we make
• 1:05:02
  a living communicate. If you see the way that we do
• 1:05:06
  our public facing communication or our stuff on YouTube, Are you
• 1:05:11
  sure we're communication professionals?
• 1:05:12
  Because we don't look like it sometimes.
• 1:05:14
  Well, that's what I was going to say, John.
• 1:05:16
  I think it's like we need to not the same thing that AI is doing
• 1:05:19
  in terms of PR, but like we need a voice.
• 1:05:22
  Like we need, like we don't use our voice other than to like to
• 1:05:28
  convey a message into another language, but we don't use our
• 1:05:30
  own thoughts necessarily. And we need to start doing that
• 1:05:33
  and start showing people that like with this podcast or with
• 1:05:37
  other tools that like, we need to tell people why we're
• 1:05:42
  important. Can I implore my fellow
• 1:05:45
  interpreters? It's.
• 1:05:46
  Possible to have a video with more than one camera angle.
• 1:05:51
  It's also possible to have a video where you don't just sit
• 1:05:53
  in a chair and talk. Yes, I know that's what we're
• 1:05:55
  doing right now. But it's possible to have a
• 1:05:57
  video where you actually physically move.
• 1:06:01
  Your phone has a good enough camera that if you take it
• 1:06:04
  outside, it will not disintegrate.
• 1:06:07
  You're allowed to be creative in the way that you're presented
• 1:06:11
  interpreting and the video with the PowerPoint slide stopped
• 1:06:16
  being interesting and about 2006 yeah so let's our per game about
• 1:06:22
  8 does I was surprised at how straightforward it was to learn
• 1:06:26
  to be more creative on YouTube I I I broke the habit of a
• 1:06:30
  lifetime I started. So my early.
• 1:06:31
  Insight, interpreting videos are all me staying at a camera.
• 1:06:35
  I've got notes. There's no script, I'm just
• 1:06:36
  gonna talk to the camera. And they were all mostly in the
• 1:06:41
  interpreters. An interpreter said, how much do
• 1:06:44
  you like them? And I got like a couple 100
• 1:06:45
  views and then I looked at the videos I was watching on YouTube
• 1:06:49
  and I was watching Veritasium before it got sold to Private
• 1:06:54
  capital as watching Smarter Everyday.
• 1:06:56
  I was watching Alex Steele. And I realized that what these
• 1:06:59
  people had in common is they were super enthusiastic and they
• 1:07:02
  were showing you things from an an unexpected angle.
• 1:07:05
  There's a channel called stand up Mass.
• 1:07:07
  The guys have 4 maths teacher who will pay he every 14th of
• 1:07:12
  March they have Pi day and he finds a new way to measure π.
• 1:07:18
  One day he got like hundreds of pies and put them on the ground
• 1:07:20
  in a circle and used, and it was fun to watch because he was
• 1:07:23
  having fun with his profession. Yeah, there's maybe no way to
• 1:07:28
  have pies. And interrupting if someone.
• 1:07:30
  Can make an interesting video about pies.
• 1:07:32
  I'd love to. Well, The thing is, in an
• 1:07:34
  interpretation, it's when. We're work.
• 1:07:37
  It's I think it doesn't come as a reflex for a lot of
• 1:07:40
  interpreters because the meeting isn't about us.
• 1:07:43
  The meeting is about other people.
• 1:07:45
  And as you said earlier, we're we're more often than not thrown
• 1:07:49
  in a situation where thrown in the middle of the story.
• 1:07:52
  And other interpreters, perhaps I mean like dozens or hundreds
• 1:07:57
  of other interpreters before us were also part of this middle of
• 1:08:01
  this story where, you know, where one more interpreting team
• 1:08:05
  that goes there and the meeting is in about us.
• 1:08:10
  So it's really hard for interpreters to say, you know
• 1:08:13
  what, hey, listen, I'm more than just a parrot that repeats into
• 1:08:19
  another language what someone else someone else says.
• 1:08:22
  Smart parrot. That's the.
• 1:08:25
  Part. That's the.
• 1:08:27
  Newest Acoustic men would be an even bigger one.
• 1:08:29
  The secret to that is we don't make the videos about us be so,
• 1:08:36
  you know the stories that we share around the fireside at the
• 1:08:39
  conference. Yeah, I invited.
• 1:08:41
  I wanted to invite. No name someone.
• 1:08:43
  And that someone said, oh, I don't want it to be about me.
• 1:08:48
  And I'm like, you're missing the point.
• 1:08:51
  Yes, it has to be about you. Yes.
• 1:08:55
  But I'm like, whatever. But the stories.
• 1:08:57
  You shared around the. Fireside the stories.
• 1:08:59
  You shared at conferences, we know how to share stories
• 1:09:02
  without breaching client confidentiality.
• 1:09:04
  So, so my my favorite story that I tell all the time, I know I
• 1:09:08
  can't even say what the client make because if I say what the
• 1:09:11
  client make, you can narrow it down to two companies.
• 1:09:14
  So I can't ever see what that client make rhymes with.
• 1:09:20
  They're just joking. Like literally if if I even.
• 1:09:24
  Told you what sector they were and you can start narrowing it
• 1:09:26
  down very because there are some sectors where there aren't that
• 1:09:29
  many people in that sector yeah, it's like so so I know I have to
• 1:09:32
  be very. Astute and I.
• 1:09:34
  Think twice in twice or three times.
• 1:09:36
  And Doctor Downey, Mr. Stick, I say I've changed some of the
• 1:09:39
  details for confidentiality. We think about that if we're on
• 1:09:44
  YouTube. And we go, well, this is going
• 1:09:45
  to be public. I have to be anonymised enough
• 1:09:47
  that the client can't recognise it, because if the client can't
• 1:09:51
  recognize it, no one else can. But what's this story about?
• 1:09:55
  This story isn't about, you know, Ohe, Yeah.
• 1:09:57
  Once interpreted, for someone who looked after pandas, it's
• 1:10:00
  never about the industry. It's never about that.
• 1:10:04
  I helped the client sell. 3010 cans.
• 1:10:06
  That's not what it's about, is no if the if the heart of every
• 1:10:10
  interpreting story is people are in a room together really
• 1:10:14
  wanting this thing to happen or this thing to not happen or
• 1:10:18
  really wanting to understand what this person has to say
• 1:10:20
  about. Yeah, but you can't because
• 1:10:24
  there's a language barrier, which doesn't matter 99% of the
• 1:10:27
  time because they're quite happy.
• 1:10:28
  You know, I joke about, you know, Missoula Baton can go on
• 1:10:31
  strike anytime he wants and doesn't need any English.
• 1:10:33
  It's fine. Yeah, but when Missoula Baton
• 1:10:35
  meets Mr. Stick and they want to negotiate car parks, well, they
• 1:10:41
  needed an interpreter. Yeah, I bet.
• 1:10:44
  What's the entire thing about the entire things not about the
• 1:10:46
  car parks, the Internet things about what happened in this
• 1:10:49
  meeting that made it work. And from that you can easily do
• 1:10:54
  a YouTube video. You can easily do YouTube videos
• 1:10:56
  on common questions that you have from clients, like why they
• 1:10:59
  need prep or why do you need us to book your two weeks in
• 1:11:02
  advance. You can easily do YouTube videos
• 1:11:05
  on funny things that clients say that are only funny if you're an
• 1:11:08
  interpreter and you can explain why it's funny.
• 1:11:10
  There's a million shorts to be made about interpreting.
• 1:11:13
  You know, I'm sure every interpreters had the thing of
• 1:11:16
  there's a rule that I haven't done a video on yet.
• 1:11:18
  Open open season or someone to take this idea of the fact that
• 1:11:21
  an interpreting would realize that the more expensive the suit
• 1:11:24
  were wearing, the more stating the buffet at the conference.
• 1:11:27
  That's just a rule that just has to happen.
• 1:11:30
  Or, you know, if you're, if you've only brought one suit
• 1:11:33
  with you, all of the desserts have high pressure cream in them
• 1:11:38
  and you know that as soon as you bite and you're gonna go.
• 1:11:42
  Are you that? Happened, but these aren't
• 1:11:44
  relatable human. Experiences that aren't just
• 1:11:46
  about interrupting, but we can tell stories in an anonymous
• 1:11:51
  way, in an entertaining way. There's a a Weedy media Perez, I
• 1:11:55
  think, I think he's been on the podcast.
• 1:11:56
  Yeah, yeah, we have her Instagram videos are great and
• 1:11:59
  she doesn't have to do. Much, I think she.
• 1:12:00
  Just has like training theme music and the words of the music
• 1:12:05
  say something about the interesting.
• 1:12:07
  It's like you just look at the YouTube channels or the
• 1:12:11
  Instagram rules that you're watching and go, what does,
• 1:12:14
  what's the interpreting version of that?
• 1:12:17
  And my favorite series recently was Ryan Trahan and his wife
• 1:12:20
  doing 56 and 50 days to raise money for Kids Research
• 1:12:24
  Hospital. And every video was identical in
• 1:12:27
  layout, but it was brilliant every time.
• 1:12:30
  Well, and you think? Well, what's the?
• 1:12:33
  Interesting version 50 interrupt.
• 1:12:34
  Doesn't 50 these if we can't find 15,050 days and we have to
• 1:12:39
  go a day without coffee, you know, you think creatively about
• 1:12:43
  what are you already watching? How does that engage with you?
• 1:12:47
  What's the interesting version of Game of Thrones?
• 1:12:48
  I don't even watch Game of Thrones.
• 1:12:49
  But you know, what's Game of Thrones interpreting style?
• 1:12:52
  Is it Game of Booths? I don't know.
• 1:12:54
  But you know what, what content are you watching?
• 1:12:56
  What's the interpreting take on that or the interpreting parody?
• 1:13:00
  And people will watch because they go.
• 1:13:02
  I had someone come up there once and say I did a talk once, a
• 1:13:06
  business event and this lady came up to me at the end and
• 1:13:08
  said oh I didn't realize interpreting was interesting,
• 1:13:16
  job done. Yeah, well, we're going to do
• 1:13:18
  something similar to. The guy who?
• 1:13:20
  Does the hot wings and they get hotter and hotter every time,
• 1:13:23
  but we didn't know how to ship wings to Scotland.
• 1:13:26
  So just get the speaker getting faster and faster and faster.
• 1:13:30
  There you. Go eventually.
• 1:13:33
  Then there's a there's an exciting keep acquire engineer.
• 1:13:36
  There you go, we'll have a. They are but.
• 1:13:38
  What one of the one? Of the things I'd like to do is
• 1:13:43
  to have found interpretation, actually move on location and
• 1:13:48
  meet people directly. There are some things that we're
• 1:13:51
  thinking about right now, but I I think that some people are
• 1:13:55
  calling you in the back and missed the mute button by three
• 1:14:01
  seconds. That's OK, my final
• 1:14:04
  interrupting. Is on this field.
• 1:14:05
  This just shows that we're. All human here so.
• 1:14:09
  But it's it's I, I did say that. You know you've been interpreted
• 1:14:13
  too long when you wondered if you can get a mute button that
• 1:14:15
  works on people. Sometimes you want a mute button
• 1:14:20
  on some people. Yeah, sadly the mute button only
• 1:14:23
  mutes the interpreters it. Never mutes the speakers.
• 1:14:27
  That's why I'm sort of jealous of people who can turn on their
• 1:14:29
  hearing aids. On and off, you know, it's like
• 1:14:33
  I don't need to hear this guy. OHS OK Can I can I put a thing
• 1:14:36
  out that I would? Like to petition that Apple
• 1:14:39
  Airpods be banned as microphones.
• 1:14:41
  It interpreted meetings. Oh my God, yes.
• 1:14:43
  Thank you. Yes.
• 1:14:45
  Yeah. Because.
• 1:14:46
  Because people think any. Bluetooth Bluetooth headset
• 1:14:50
  basically people think. They look so cool with that
• 1:14:52
  player pods. And they, they, they look like,
• 1:14:55
  kind of like a catalogue model until they unmute and then they
• 1:14:59
  sound like a Dalek from Doctor Who or they.
• 1:15:04
  Sound like, um, you know when mobile phones.
• 1:15:07
  Were first out in every call were really crackly yeah it's
• 1:15:11
  that it's probably of course stop yeah, let's interpret.
• 1:15:21
  That let's see how that works said that through AI so one time
• 1:15:26
  I was. Trying to explain to a client
• 1:15:27
  and. The client said, why are Airpods
• 1:15:32
  bad? And I said, well, do you think I
• 1:15:37
  can do my job as an interpreter? And if if I can't hear the
• 1:15:40
  information, like, what do I need to hear?
• 1:15:44
  Right? Well, if I can't hear, I can't
• 1:15:47
  do it, right? OK, right.
• 1:15:51
  Yeah, those things, hearing is essential.
• 1:15:54
  Those explanations work. I'm I'm planning at some point
• 1:15:56
  I. Don't have airports.
• 1:15:58
  At some point I'm going to do the comedy version of what you
• 1:16:01
  would like. I'm guilty.
• 1:16:02
  I'm guilty, but it's like what you look like.
• 1:16:04
  Versus what you sound like people don't know.
• 1:16:06
  Until we tell them. And it's like, I'm amazed I've
• 1:16:09
  come across people who think that we can just walk into any
• 1:16:11
  meeting and interpret. I'd love to be able to.
• 1:16:14
  I know that interpreters who have to do that because of their
• 1:16:16
  job. And they must be a better
• 1:16:18
  interpreter than me because I, I love having the prep materials
• 1:16:22
  in front of me. And I find I feel much more able
• 1:16:25
  to do well when I know right, we're going to be talking about
• 1:16:28
  this. Although although the fourth
• 1:16:33
  rule of interpreting is never trust an agenda because.
• 1:16:38
  Because we all know that there's going to be.
• 1:16:41
  Two things about never trusting an agenda, and I think we need
• 1:16:43
  to go soon, but never, never trusting an agenda.
• 1:16:45
  One of them is it's the list of all the speeches that will run
• 1:16:48
  long, be moved or not happened that day.
• 1:16:50
  That's what an agenda is. And two, if you look on the
• 1:16:54
  agenda and see those 8 speeches during a day, if you have seven
• 1:17:00
  sets of notes from that agenda and there's one speech missing,
• 1:17:05
  it will always be the most technical involved one,
• 1:17:11
  including quotes from legislation accounts from 1989.
• 1:17:16
  And if it's French, a quote from a dead philosopher, you know it
• 1:17:19
  will, it will all be in there. And that will be the one you
• 1:17:22
  didn't get the notes for. And the speaker who apologizes
• 1:17:24
  for sending you their notes is going to be the best speaker to
• 1:17:28
  interpret that this town agenda works.
• 1:17:30
  Yeah, on that note, yes, on. That note everyone I recommend.
• 1:17:35
  So interpreters versus machines, this is the hard copy and it's
• 1:17:46
  sequel or follow up this. Is Mr. Downey.
• 1:17:49
  And Mr. Stick guide to the future of interpreting available
• 1:17:55
  now. The link will be in the
• 1:17:57
  description below. I encourage every interpreter
• 1:18:00
  watching this podcast to read this books.
• 1:18:04
  I know we digressed a lot during the podcast.
• 1:18:07
  We talked about all kinds of things and this is what we do on
• 1:18:09
  a podcast. But when it comes to the book,
• 1:18:12
  it's really, really tight and it's not a very long read.
• 1:18:16
  But actually I did learn a lot and it really opened my eyes.
• 1:18:21
  And actually, to be honest with you, it made me feel more
• 1:18:26
  confident about the future of our industry in a time where I
• 1:18:32
  feel where people are overly pessimistic.
• 1:18:37
  And I think it's important to inject some hope not but more
• 1:18:42
  and more than just hope the fun and show that we're here to stay
• 1:18:48
  and don't believe the hype. Yeah.
• 1:18:51
  And and Jonathan, thank you so much for coming on.
• 1:18:53
  It's been. A pleasure again and like like
• 1:18:57
  Alan says, we digress a lot, but we do it and it's fun and we
• 1:19:01
  learn along the way. It was great to be back, to be
• 1:19:04
  challenged again and I have to see all the artwork.
• 1:19:06
  And Doctor Diane and Mr. Sticking to the future of
• 1:19:08
  interpreting was hand drawn by me.
• 1:19:11
  Yeah, it's been done by AI. Would probably be better.
• 1:19:14
  What about the cover? What about the cover?
• 1:19:16
  That cover was designed by A. Professional Designer.
• 1:19:18
  Routledge. Ah, yes, good.
• 1:19:21
  So the Routledge, because all professional editors, designers,
• 1:19:24
  Doctor Daniel, Mr. Sticks was with me doing it.
• 1:19:27
  There is also a large print version which I will leave in
• 1:19:30
  the link. But yeah, thank you for having
• 1:19:31
  me back on again. It was good to get.
• 1:19:33
  I like some of those devil's advocate questions as well.
• 1:19:36
  And I think the the main thing that I want to leave people is
• 1:19:39
  with is be glad that you're an interpreter and actually do
• 1:19:44
  something to make your voice that because the world needs
• 1:19:46
  interpreters to stand up. I know what used to being
• 1:19:49
  invisible interpreter, but stand up, be visible and be creative
• 1:19:55
  in how you tell people about your work.
• 1:19:56
  And you never know, one of your shorts that you make on a whim
• 1:19:59
  might educate someone else's client and suddenly you make out
• 1:20:02
  the day better for other interpreters.
• 1:20:03
  It's it's happened to me once. It can happen to you.
• 1:20:06
  I'm more than happy if someone does shorts and my clients start
• 1:20:09
  behaving better that that would make my day.
• 1:20:13
  There'll be shorts of this episode, that's for sure.
• 1:20:16
  Right, everyone, I'll see you all.
• 1:20:18
  Next time for another episode. Of final interpretation podcast
• 1:20:21
  Take care, bye.