Latest / How I Work / Stop writing like a robot - ban these AI words and phrases
Transcript
- 0:00Speaker 1: You can always tell when something's been written by AI with.
- 0:05Speaker 2: Little human involvement. You know what I mean?
- 0:09Speaker 1: A LinkedIn post that starts with in today's fast paced world,
- 0:13Speaker 1: or an article that ends with it's not about X,
- 0:16Speaker 1: it's about why. A paragraph but technically fine, but somehow
- 0:20Speaker 1: feels like it was written by nobody. And that is
- 0:25Speaker 1: the uncanny valley of AI writing. And the problem isn't
- 0:30Speaker 1: just that it sounds robotic, it's that people stop reading.
- 0:34Speaker 1: The frustrating part is that this happens even when you
- 0:38Speaker 1: have done the thinking, You write the ideas, hand it
- 0:42Speaker 1: to AI for a cleanup, and it AI effies the
- 0:46Speaker 1: whole thing, flattening your voice and coding it with M
- 0:50Speaker 1: dashes until nothing sounds like you anymore. So today Neo
- 0:55Speaker 1: and I get into exactly how to stop that from happening.
- 0:59Speaker 1: We cover specific instructions you can give to AI to
- 1:02Speaker 1: keep your writings sounding human, things like mixing sentence lengths,
- 1:07Speaker 1: avoiding stiff transitions, and banning the.
- 1:10Speaker 2: Rule of three. To see what I did there, I
- 1:12Speaker 2: just choose it. We also go through.
- 1:15Speaker 1: Some of the most overused AI words and phrases to
- 1:18Speaker 1: strike from your prompts entirely and now as a little bonus,
- 1:23Speaker 1: you will find a link in the show notes to
- 1:25Speaker 1: our anti AI slot prompt, which is a ready to
- 1:28Speaker 1: use resource you can customize with your own personal pet
- 1:31Speaker 1: hates and run any content through before it goes out. Okay,
- 1:36Speaker 1: let's get on with the show. Welcome to how IAI
- 1:45Speaker 1: with me, Doctor Amantha Imber and Neo Applin, head of
- 1:49Speaker 1: Inventium AI. Each episode we share one practical way to
- 1:53Speaker 1: use AI better at.
- 1:55Speaker 2: Work and in life.
- 1:56Speaker 1: No fluff, no tech jargon, just things you can use
- 2:00Speaker 1: straight away. So much content in the world is clearly
- 2:05Speaker 1: written by AI now, and I know that it.
- 2:07Speaker 2: Really frustrates us, But like, why is it a problem?
- 2:10Speaker 1: Why should we actually be aiming to not sound like
- 2:13Speaker 1: AI when we are all using AI to create the
- 2:16Speaker 1: content and words we are putting out into the world.
- 2:19Speaker 3: Now, Yeah, I think it's really two parts. One is
- 2:22Speaker 3: don't be AI, in other words, don't just get AI
- 2:24Speaker 3: to do all your work for you. And then the
- 2:26Speaker 3: second part is don't sound like AI. And I think
- 2:29Speaker 3: because a lot of people are now doing things where
- 2:31Speaker 3: they're going like five dot points and say write me
- 2:33Speaker 3: a big article, and it might be there are five
- 2:35Speaker 3: pages that really Ai wrote all of it apart from
- 2:37Speaker 3: five dot points worth, and it just feels and smells
- 2:41Speaker 3: like AI, and it just has all the tales like
- 2:44Speaker 3: everyone knows about the M dashes. But you can just
- 2:46Speaker 3: read it and go, ah, this doesn't sound It sounds
- 2:49Speaker 3: like CHGBT. And so you can get AI to do
- 2:52Speaker 3: all the work and it sounds like AI. The other
- 2:55Speaker 3: thing is you can do most of the work and
- 2:57Speaker 3: then get AI to do a cleanup at the end,
- 3:00Speaker 3: and it will sound like AI. And so everyone's going
- 3:02Speaker 3: to look at that and go, the whole thing's written
- 3:04Speaker 3: by AI, and then they'll disregard it like they probably
- 3:07Speaker 3: should about the first one, which is just all written
- 3:10Speaker 3: by AI. So if you build some stuff that sounds
- 3:14Speaker 3: like AI, people are going to roll their eyes and
- 3:17Speaker 3: they may not read it with the same kind of
- 3:19Speaker 3: intent or kind of care that they would if it
- 3:21Speaker 3: was sounding like you. And they go, oh my god,
- 3:23Speaker 3: I Mattha right the whole thing, And I really want
- 3:25Speaker 3: to listen to everything she was said, and I really
- 3:27Speaker 3: want to read it intensely rather than I don't know
- 3:29Speaker 3: if AIS written this. Therefore I'll just skim it, or
- 3:32Speaker 3: I'll get AI to summarize it.
- 3:33Speaker 1: So the point of this episode is not to say
- 3:36Speaker 1: don't use AI to help you write things, but perhaps
- 3:41Speaker 1: don't use AI to do all the thinking and all
- 3:43Speaker 1: the content creation. And so then the point of us
- 3:48Speaker 1: helping you not sound like AI means that the work
- 3:51Speaker 1: that you do produce, that you do genuinely think about
- 3:54Speaker 1: and work with AI to refine and edit and so forth,
- 4:00Speaker 1: gets taken a whole lot more seriously rather than people
- 4:02Speaker 1: just dismissing it. So Neo, I know that in the
- 4:06Speaker 1: Inventium AI training, when we're teaching people how to use
- 4:13Speaker 1: AI for writing, you actually do go into this. So
- 4:15Speaker 1: what are some of the things that you're teaching people
- 4:18Speaker 1: to avoid sounding like AI?
- 4:21Speaker 3: The first one I will say is this, Like often
- 4:24Speaker 3: we'll get it like a paragraph and you might say, look,
- 4:26Speaker 3: I need you to clean up this paragraph AI, so
- 4:28Speaker 3: you're not getting it to write the whole damn thing.
- 4:31Speaker 3: And the first tip is telling AI don't make this
- 4:35Speaker 3: sound more complicated than the original. So IT often likes
- 4:39Speaker 3: to have these big words and be very stiff and
- 4:41Speaker 3: quite formal unless you've previously told it you know what
- 4:45Speaker 3: you like it to sound like, and things like that.
- 4:47Speaker 3: So the quickest easiest one is just saying something along
- 4:49Speaker 3: the lines of, don't make the output more complicated than
- 4:52Speaker 3: in the regional when you give it a paragraph and
- 4:54Speaker 3: ask it to critique it and clean it up and
- 4:56Speaker 3: things like that. But there are other ones that you
- 4:59Speaker 3: can add in. So the first one is the most
- 5:01Speaker 3: obvious one, but there's a bit of a catch on
- 5:03Speaker 3: how to do it. You know those M dashes that
- 5:05Speaker 3: seem to pop up everywhere, So I'll just explain for
- 5:08Speaker 3: those who aren't sure what an M dash is, and
- 5:10Speaker 3: M dash is a longer dash. This is a I
- 5:12Speaker 3: probably should get you to explain this, you're the actual
- 5:14Speaker 3: author here, but yeah, it's a proper way to write things, right.
- 5:19Speaker 1: Yes, yeah, So they're like visually they look like a
- 5:22Speaker 1: double hyphen So if you can imagine two hyphens connected,
- 5:26Speaker 1: that's what an m dash looks like. And because AI
- 5:29Speaker 1: was trained on books, if you read any book and
- 5:34Speaker 1: they're using you know, like language like and I'm just
- 5:40Speaker 1: trying to think of a sentence that has like an
- 5:42Speaker 1: M dash in it, so you know it might be
- 5:45Speaker 1: run every piece of content through this list before publishing,
- 5:51Speaker 1: and if you spot three or more, rewrite, so like
- 5:53Speaker 1: you could theoretically have an M dash before the end,
- 5:57Speaker 1: just to kind of break up the sentence.
- 5:58Speaker 2: You'll see it in any book that you're reading.
- 6:01Speaker 1: And because AI was trained on all those books, a
- 6:03Speaker 1: lot of M dashes went into the system in terms
- 6:06Speaker 1: of its knowledge, and it's.
- 6:07Speaker 3: Really hard baked in there because AI is a product
- 6:11Speaker 3: of its training, because it was trained on so many books,
- 6:14Speaker 3: then M dashes are everywhere throughout its training, and it
- 6:17Speaker 3: just pulls out M dashes an awful lot, and it's
- 6:19Speaker 3: really hard to get AI to not use M dashes.
- 6:22Speaker 3: But so rather than just say don't use M dashes,
- 6:26Speaker 3: I found that that has a very low chance of working.
- 6:29Speaker 3: Sometimes it does gish kind of Sometimes. The phrase that
- 6:32Speaker 3: I use, which has given me the best success is this.
- 6:36Speaker 3: Now it's don't use M dashes, So that is, don't
- 6:40Speaker 3: use em space dashes. And then I actually give it
- 6:44Speaker 3: an example of an M dash, So I actually paste
- 6:47Speaker 3: in an M dash within brackets, so it then knows
- 6:49Speaker 3: that's an M dash. I've been really clear about this,
- 6:53Speaker 3: and you could if you don't like short dashes either,
- 6:55Speaker 3: you could say or short dashes, and then you can
- 6:57Speaker 3: put the short dash in an inner brackets. Rely on
- 7:01Speaker 3: commas full stops or short shorter sentences instead, So you
- 7:05Speaker 3: could do things like that, or you could write it
- 7:06Speaker 3: if you're affine with the dash, you just want them
- 7:09Speaker 3: to be short. It'd be don't use M dashes, paste
- 7:11Speaker 3: in an M dash instead. I want you to use
- 7:14Speaker 3: short dashes and then paste in the short dash. So
- 7:16Speaker 3: it's up to you whether you want to ban the
- 7:18Speaker 3: dash or you just want to keep the short dash,
- 7:21Speaker 3: but yet give it first off, the words and also
- 7:24Speaker 3: the example of what the dash is you're talking about,
- 7:26Speaker 3: and I found that is better, but not one hundred
- 7:29Speaker 3: percent why it loves the M dash and it's really
- 7:32Speaker 3: hard to get it to stop doing.
- 7:36Speaker 1: What are some of the other things that you teach
- 7:38Speaker 1: in inventium dot AI programs.
- 7:42Speaker 3: These are all to undo some of the very common
- 7:46Speaker 3: things that AI does, and with AI doing in the
- 7:50Speaker 3: commonly things, it's doing it the right way. And I'm
- 7:53Speaker 3: doing your quotes here because the right way to write
- 7:56Speaker 3: is the very formal way to write, and it's been
- 7:58Speaker 3: trained on the right way to write by a whole
- 8:00Speaker 3: bunch of books and sources, and also you know the
- 8:02Speaker 3: guides on how to write well, and so therefore you
- 8:05Speaker 3: know if you're a native English speaker and you're you know,
- 8:08Speaker 3: you've done a lot of writing, chances are you more
- 8:11Speaker 3: get caught of these AI detectors and things like that.
- 8:15Speaker 3: So it's almost like putting the human back into it,
- 8:18Speaker 3: making it a little bit messier than the perfect way
- 8:21Speaker 3: to write. Okay, So with that first one's the second
- 8:24Speaker 3: one is mix sentence lengths. And the reason for this
- 8:27Speaker 3: is AI generally has the same sentence length again and
- 8:30Speaker 3: again and again, and so when you and I read this,
- 8:34Speaker 3: we may not actually pick up that these sentence lengths
- 8:37Speaker 3: are the same, but we just go this doesn't feel
- 8:39Speaker 3: like a human so we get that kind of like
- 8:41Speaker 3: the uncanny valley of writing. So say mix sentence lengths,
- 8:45Speaker 3: these short ones for punch. He's longer ones when you
- 8:48Speaker 3: need detail, And I found that one does pretty well.
- 8:51Speaker 3: So if you remember hending Away famous for short, punchy sentences,
- 8:56Speaker 3: so get some punchy things in there, get some detailed
- 8:58Speaker 3: ones that mixes it up and you and I reading
- 9:00Speaker 3: it go okay. Another one is the language that AI uses.
- 9:05Speaker 3: It can try to be the smartest AI in the
- 9:07Speaker 3: room and can like to use those long words as
- 9:11Speaker 3: part of that don't make the thing more complicated. So
- 9:14Speaker 3: you could do something like this, use playing language at
- 9:17Speaker 3: a ninth at about a ninth grade and reading level,
- 9:20Speaker 3: avoid jargon and buzzwords. Now that the ninth grade thing
- 9:25Speaker 3: or the fourteen year old type brief is common along
- 9:32Speaker 3: in industry practice, I understand when the newspaper or pro
- 9:34Speaker 3: writing and things like that, that's the kind of goal to
- 9:37Speaker 3: get to. Why because everyone can just read that pretty simply,
- 9:42Speaker 3: pretty easily. It's not really difficult to read, and you're
- 9:44Speaker 3: not going to pull out university style words that are
- 9:46Speaker 3: only a handful of people going to understand. So that's
- 9:49Speaker 3: why you do that. The other ones are sometimes it
- 9:54Speaker 3: deals with the way that sentences is constructed. Particularly start
- 9:57Speaker 3: of sentences can be a bit stiff. So one of
- 10:00Speaker 3: the things is it's fine to start a sentence with
- 10:02Speaker 3: and but also although you should do this sparingly, so
- 10:06Speaker 3: occasionally we will actually start a sentence with a handbuilt
- 10:09Speaker 3: but or so that's not the right way to do it.
- 10:10Speaker 3: I'm doing air quotes again, but humans do that, and
- 10:14Speaker 3: so therefore let's let AI do it. Not do it
- 10:16Speaker 3: all the time, but we'll let it do it. The
- 10:18Speaker 3: other one about starting sentences is avoid stiff transitions such
- 10:22Speaker 3: as additionally and moreover. Who's start human says moreover blah
- 10:28Speaker 3: blah blah blah blah. You know, you feel like you're
- 10:31Speaker 3: like a stuffy person in the eighteen hundreds with a
- 10:33Speaker 3: pipe or something. So I avoid stiff transitions such as
- 10:36Speaker 3: additionally and moreover. Now this is a weird one, but
- 10:41Speaker 3: I find this one. I used to like, but now
- 10:45Speaker 3: I hate that AI does. Often it'll frame things in
- 10:50Speaker 3: a particular way near the end of a sentence, end
- 10:53Speaker 3: of a paragraph. It'll say half bunch of explanation stuff,
- 10:57Speaker 3: and I want to say, it's not this. It's so it
- 11:02Speaker 3: maybe like, it's not just a good way of writing,
- 11:06Speaker 3: it's the way to communicate something.
- 11:10Speaker 2: Like that, right, AI.
- 11:13Speaker 3: It does this a lot, and actually was at the
- 11:15Speaker 3: start when it was coming, I was, Oh, that's really clever.
- 11:17Speaker 3: It does that really well. I like, but it over
- 11:20Speaker 3: uses that mechanism, and so therefore I actually had to
- 11:23Speaker 3: google to find out what that mechanism was called. It's
- 11:25Speaker 3: called corrective antithesis. So I say, never use corrective antithesis,
- 11:31Speaker 3: And then in brackets, given an example, I say, for example,
- 11:34Speaker 3: not X, but why and then that certainly cuts that down.
- 11:39Speaker 3: And there's one more if I can throw it in there,
- 11:42Speaker 3: which is don't lean on the rule of three. Now,
- 11:45Speaker 3: the rule of three is generally you give examples. You
- 11:47Speaker 3: can give three examples, and it kind of makes you know,
- 11:50Speaker 3: it reads nicely. But AI over uses the rule of three.
- 11:56Speaker 3: It over uses that in so many things, and so
- 11:59Speaker 3: I say, don't lean on the rule of three. Use
- 12:01Speaker 3: one or two items, or use a longer list if
- 12:04Speaker 3: it strengthens the point. So you can do one or two,
- 12:07Speaker 3: or you can do five. You know, up to you,
- 12:10Speaker 3: but don't always use rule of three, because it's that consistency,
- 12:14Speaker 3: like the sentence lengths being the same, the rule of
- 12:16Speaker 3: three being the same. It's not this, it's that being
- 12:19Speaker 3: the same that feels very robotic. Because robots are really
- 12:22Speaker 3: good at doing things the same ais really good things
- 12:24Speaker 3: the same. Humans we're messier, and so we're used to
- 12:27Speaker 3: seeing things that a little bit messier. And it's also
- 12:29Speaker 3: easier and nicer for us to read because it kind
- 12:31Speaker 3: of feels more like us. And so this is just
- 12:34Speaker 3: a whole bunch of points you can put in if
- 12:36Speaker 3: like you've got this as a standard prompt, all those
- 12:39Speaker 3: what have I got one to three, four, over, six, seven, eight,
- 12:41Speaker 3: seven or eight dot points? You can just drop them
- 12:44Speaker 3: in and then say, when you rewrite my paragraph because
- 12:46Speaker 3: I needed to be cleaned up, don't do these things.
- 12:49Speaker 3: That's what you're trying to tell them.
- 12:51Speaker 2: Okay, I love all those.
- 12:52Speaker 1: Now what I did ne in preparation for this episode
- 12:56Speaker 1: is I went into all the llms that I use
- 13:04Speaker 1: and like I love comparing the output, so I went
- 13:08Speaker 1: into chatpt Claude, Gemini and also Perplexity, and I gave
- 13:14Speaker 1: it all the same prompt, which was something along the
- 13:16Speaker 1: lines of.
- 13:19Speaker 2: Tell me what are the fifty.
- 13:22Speaker 1: Most overused or over indexed words and phrases that AI
- 13:28Speaker 1: likes to use, and I think I suggested to look
- 13:31Speaker 1: at LinkedIn for some you know, some examples. And then
- 13:37Speaker 1: what I did with the output that I got from
- 13:39Speaker 1: each of those four models, I then put them all
- 13:44Speaker 1: into the one model Claudes generally the one that I
- 13:46Speaker 1: default too, and I said, give me a summary of
- 13:49Speaker 1: the ones that are appearing most frequently across these four outputs.
- 13:54Speaker 1: And so with that, I've then created a bit of
- 13:58Speaker 1: a mega prompt that people can use that I've called
- 14:01Speaker 1: the anti ai slop prompt and if you want that,
- 14:05Speaker 1: I will link to it in the show notes. But
- 14:07Speaker 1: I'm just going to go through a few things just
- 14:09Speaker 1: for some additional advice for people that are trying to
- 14:13Speaker 1: un ai aify their writing so often. And these are
- 14:19Speaker 1: grouped into some categories. So there's the first category is
- 14:22Speaker 1: empty opener phrases. So you will have seen things like
- 14:27Speaker 1: this in today's fast paced world, in an ever changing
- 14:30Speaker 1: landscape when it comes to.
- 14:32Speaker 2: Blah, picture this black like.
- 14:34Speaker 1: These are all sounding really familiar because these are how
- 14:36Speaker 1: a lot of certainly a lot of LinkedIn post starts start. Now,
- 14:41Speaker 1: then we've got a category of fake profound hooks, which
- 14:45Speaker 1: I think, you know, it's not X, it's why is
- 14:47Speaker 1: certainly one of those. Here's the truth, which is one
- 14:52Speaker 1: of my pet hates, or sometimes he's the uncomfortable truth,
- 14:55Speaker 1: he's the inconvenient truth. I hate that one because I
- 14:59Speaker 1: use that language here the truth or here's the thing.
- 15:02Speaker 1: I actually use that a lot in my writing. And
- 15:04Speaker 1: what is really frustrating, as someone who was writing books
- 15:07Speaker 1: before the age of AI is that I feel like
- 15:10Speaker 1: I now have to change my writing style, which was
- 15:14Speaker 1: you know, certainly in some ways similar to what AI
- 15:17Speaker 1: produces to you know, just added like I used m
- 15:20Speaker 1: dashes very often, but now I will lean more towards
- 15:23Speaker 1: brackets and semi collins for like little asides or quips
- 15:29Speaker 1: as my editor calls them, which is very frustrating when
- 15:32Speaker 1: you have to change your own writing style so it
- 15:34Speaker 1: doesn't look like AI. And we've got lots of corporate buzzwords.
- 15:38Speaker 1: Probably the most common is delve. I was hearing some
- 15:42Speaker 1: research actually that was done in twenty twenty three where
- 15:47Speaker 1: academic authors that were most prolific in twenty twenty three
- 15:52Speaker 1: in terms of getting things published over indexed heavily on
- 15:55Speaker 1: the word delve. I don't know if you've heard that research.
- 15:59Speaker 1: I heard Ethan like sharing that recently, which made me laugh.
- 16:03Speaker 3: I didn't hear that researcher. I'm a bit sad about
- 16:05Speaker 3: delve because I use delve as well, but AI uses
- 16:09Speaker 3: it more. And so you may add a bunch of words.
- 16:12Speaker 3: Can we also talk about other words we may want
- 16:14Speaker 3: to add to that banned list?
- 16:17Speaker 2: For sure? For sure? Give me like three.
- 16:19Speaker 3: More quietly is another one that I was going.
- 16:24Speaker 2: To say quietly.
- 16:26Speaker 1: So the next category actually is what AI has termed
- 16:29Speaker 1: hollow adjectives and nouns and adjectives, and if you don't
- 16:34Speaker 1: remember from your English class, and adjective is a descriptive
- 16:38Speaker 1: word for a noun, So it describes a noun which
- 16:42Speaker 1: is like an object or a person, a person, place
- 16:46Speaker 1: or thing. I think is how my English teacher probably
- 16:48Speaker 1: would have described it. Quietly is so overused. And again
- 16:54Speaker 1: that really bothers me because that is a word that
- 16:56Speaker 1: I would have used, but now I feel like I
- 16:58Speaker 1: have to strip it out mostly from my writing.
- 17:00Speaker 2: Okay, what are the words near this?
- 17:03Speaker 3: Ultimately, I would recommend you build your own list of
- 17:06Speaker 3: words that you don't want AI to use, because there's
- 17:08Speaker 3: some there that AI will have used, like Delvin quietly,
- 17:11Speaker 3: but there may be others that IT would use that
- 17:14Speaker 3: I find words but you would never use. So I
- 17:17Speaker 3: use the example of like buzzword bingo, so like synergistic. Right,
- 17:21Speaker 3: if AI uses the words synergistic and you would never
- 17:24Speaker 3: use it, then put a list together and add to
- 17:27Speaker 3: that list, but sudogistic in there, maybe put delve in there,
- 17:29Speaker 3: maybe put quietly in there, and go, don't ever use
- 17:32Speaker 3: these words, And so build up that list and enhance
- 17:34Speaker 3: that over time as AI uses words that just great
- 17:37Speaker 3: on you, because if you've been grading on you, they
- 17:39Speaker 3: don't sound like you.
- 17:41Speaker 2: So hopefully.
- 17:42Speaker 1: I mean, this is so much to take in, as
- 17:44Speaker 1: I said, for those that are listening and like, yeah,
- 17:47Speaker 1: I really, I really need to improve my writing output
- 17:50Speaker 1: because I could relate to a lot of those things,
- 17:53Speaker 1: and I think about my writing. Yes, I'm using a
- 17:55Speaker 1: lot of the things that Neo and a manter I
- 17:56Speaker 1: have talked about. Go to the show notes. There is
- 18:00Speaker 1: a link to get our anti AI slopped prompt, which
- 18:04Speaker 1: I highly recommend you customize.
- 18:07Speaker 2: So it's a lengthy prompt.
- 18:09Speaker 1: It includes a lot of what we've spoken about today,
- 18:12Speaker 1: but do read it and put in your own personal
- 18:16Speaker 1: pet peeves so that you can customize it and run
- 18:19Speaker 1: any content that you create or create with AI through
- 18:23Speaker 1: this prompt. It will improve your work a lot. And
- 18:26Speaker 1: also I would say, please share this episode with maybe
- 18:30Speaker 1: some people that you know that are perhaps producing a
- 18:33Speaker 1: bit of AI slot, but they are actually doing the
- 18:35Speaker 1: thinking themselves, which is the worst combination where the human
- 18:38Speaker 1: is doing the majority of the work and AI is
- 18:41Speaker 1: then going over it and AI are fying it and
- 18:44Speaker 1: completely destroying that person's credibility, so do share this episode
- 18:48Speaker 1: with those that need it. How i AI was hosted
- 18:53Speaker 1: by me Amantha Imber and Neo Applan. A big thank
- 18:56Speaker 1: you to Martin Imba who does our sound editing and
- 18:58Speaker 1: Jem Rubio for production support, and thank you to John
- 19:02Speaker 1: Kilby who composed the theme music