Latest / Azadeh and Guy on AI / Not All AI Is GenAI!

Transcript

Not All AI Is GenAI! · Azadeh and Guy on AI

• 0:00
  SPEAKER_00: Hi.
• 0:01
  SPEAKER_00: Hello.
• 0:02
  SPEAKER_00: How are you?
• 0:02
  SPEAKER_00: I'm good.
• 0:03
  SPEAKER_00: How are you?
• 0:04
  SPEAKER_00: I am well.
• 0:05
  SPEAKER_00: This is Episode 9 when you said that.
• 0:07
  SPEAKER_00: Every time I'm surprised.
• 0:09
  SPEAKER_00: I don't know why.
• 0:10
  SPEAKER_02: Yes.
• 0:11
  SPEAKER_02: Yeah.
• 0:12
  SPEAKER_00: But this is Episode 9 of...
• 0:15
  SPEAKER_02: Azadeh.
• 0:16
  SPEAKER_00: And Guy.
• 0:17
  SPEAKER_02: On AI.
• 0:17
  SPEAKER_00: Yes.
• 0:18
  SPEAKER_00: And so today we're really talking about two topics that we thought about.
• 0:23
  SPEAKER_00: It's a little bit different than the last few episodes.
• 0:25
  SPEAKER_00: We wanted to spice it up and make it interesting.
• 0:29
  SPEAKER_00: So this time,
• 0:30
  the first topic is I really want to talk about highlighting the
• 0:36
  difference between
• 0:37
  SPEAKER_00: AI and Gen AI.
• 0:39
  SPEAKER_00: So when we say Gen AI, we mean
• 0:42
  SPEAKER_02: Generative artificial intelligence.
• 0:44
  SPEAKER_00: Yes.
• 0:45
  SPEAKER_00: And that is the thing that the chat GPTs, the clods, the mistrows.
• 0:51
  SPEAKER_00: Yeah.
• 0:51
  SPEAKER_00: All of those, that's what they do.
• 0:53
  And that's very different than other AIs because we've had AI
• 0:58
  for a long time.
• 0:58
  SPEAKER_01: Mm-hmm.
• 0:59
  For example,
• 1:00
  we all know that the algorithms for Netflix has an element of
• 1:06
  AI.
• 1:06
  SPEAKER_00: It's got machine learning.
• 1:08
  SPEAKER_00: It learns something about you.
• 1:09
  SPEAKER_00: It also has something general.
• 1:11
  SPEAKER_00: And then it gives you a result.
• 1:13
  SPEAKER_00: This is what we think you want to watch.
• 1:15
  SPEAKER_00: And for example,
• 1:17
  SPEAKER_00: when we talk about,
• 1:18
  SPEAKER_00: for example,
• 1:19
  SPEAKER_00: MRI and other software,
• 1:22
  SPEAKER_00: we use machine learning to interpret the results.
• 1:26
  SPEAKER_00: But that is very different.
• 1:29
  SPEAKER_00: from generative AI.
• 1:31
  Generative AI is that thing that hallucinates and gives you
• 1:36
  this human-like
• 1:37
  SPEAKER_00: responses and is biased and is all of these things we're worried about.
• 1:41
  And the problem is,
• 1:43
  for me at least,
• 1:44
  people are putting it together and they talk about AI.
• 1:48
  And the problem is then we're confused exactly what are we
• 1:53
  talking about.
• 1:54
  So my request to everybody out there is be very precise about
• 1:59
  what you're using
• 2:00
  SPEAKER_00: because AI is so huge right now.
• 2:03
  SPEAKER_00: And it's infiltrating everything.
• 2:06
  And I think AI has been around for a much longer time than
• 2:09
  generative AI.
• 2:10
  SPEAKER_00: Yeah.
• 2:11
  SPEAKER_00: Generative AI is really a specific case of AI.
• 2:14
  SPEAKER_00: It's a specific technology.
• 2:16
  SPEAKER_00: And so what I'm saying is maybe in all of these cases,
• 2:20
  we need to be a little bit more specific about what is it that
• 2:25
  we mean and not just
• 2:27
  SPEAKER_00: say AI.
• 2:28
  Because even commercially,
• 2:30
  if you look out there,
• 2:31
  everybody's advertising that they're using AI.
• 2:34
  SPEAKER_01: Yeah.
• 2:34
  SPEAKER_00: It's not always generative AI.
• 2:36
  SPEAKER_00: And sometimes I'm not sure it's AI at all.
• 2:39
  It's just a marketing world at this point because everybody's
• 2:43
  interpreting it as
• 2:44
  SPEAKER_00: new,
• 2:45
  SPEAKER_00: as smarter,
• 2:46
  SPEAKER_00: as better.
• 2:46
  SPEAKER_02: Yeah.
• 2:50
  The simple message from my perspective is be specific about
• 2:53
  what you're talking about.
• 2:54
  SPEAKER_00: That's why in most of the writing we're doing,
• 2:57
  SPEAKER_00: with very few exceptions,
• 3:00
  SPEAKER_00: we talk about generative AI or Gen AI for short,
• 3:04
  SPEAKER_00: to make sure that people know what we mean and what we don't mean.
• 3:07
  SPEAKER_02: Exactly, yeah.
• 3:09
  SPEAKER_00: And my other favorite example,
• 3:12
  SPEAKER_00: just to stretch this a little bit longer,
• 3:13
  SPEAKER_00: is I'm really,
• 3:16
  SPEAKER_00: for example,
• 3:17
  most of the pictures we take now with our phones have an
• 3:19
  element of an AI and those
• 3:21
  SPEAKER_00: kind of things.
• 3:22
  SPEAKER_00: Again, it is still an AI, but it's very different.
• 3:26
  SPEAKER_00: It works differently.
• 3:27
  SPEAKER_00: It does different things.
• 3:29
  SPEAKER_00: And the interaction with it is inherently different.
• 3:32
  SPEAKER_00: So, again...
• 3:34
  SPEAKER_00: It's really important to talk about what it is specifically we mean.
• 3:38
  SPEAKER_00: And let's move to the next topic because I'm really excited about it.
• 3:41
  SPEAKER_02: Yeah.
• 3:42
  SPEAKER_02: And today we are talking about one of our research studies.
• 3:47
  SPEAKER_02: It's a kind of collaborative research project.
• 3:50
  We are doing this research with one of our collaborators in
• 3:56
  University of Maryland.
• 3:57
  SPEAKER_02: Yeah.
• 3:58
  SPEAKER_02: Yeah.
• 3:59
  And for that kind of a study,
• 4:01
  for the data collection part,
• 4:03
  we are using an online survey.
• 4:06
  SPEAKER_02: We have distributed the survey across Nebraska.
• 4:09
  SPEAKER_02: And also we try to find collaborators in other states.
• 4:13
  SPEAKER_00: In Maryland and Tennessee and a few other places.
• 4:16
  SPEAKER_00: So if you are interested, by the way.
• 4:18
  SPEAKER_00: we can send you the link.
• 4:19
  SPEAKER_00: Just drop us a note on Substack or in our emails,
• 4:25
  SPEAKER_00: amgtraining2.unl.edu,
• 4:28
  and we will be sure to send you a link because we are excited
• 4:31
  to hear from as many
• 4:32
  SPEAKER_00: language arts teachers as possible.
• 4:35
  So if you're teaching English,
• 4:36
  if you're teaching composition,
• 4:38
  if you're teaching elementary...
• 4:40
  SPEAKER_00: You teach everything, including language arts.
• 4:42
  SPEAKER_00: If you're teaching English language learners, we'd love to hear from you.
• 4:47
  This is a moment in time where we can find out a lot of new
• 4:50
  things from teachers,
• 4:52
  SPEAKER_00: and that's what we're doing here.
• 4:54
  SPEAKER_02: So we have distributed a survey,
• 4:57
  SPEAKER_02: and so far we have collected a couple of responses,
• 5:00
  SPEAKER_02: more than a couple of responses,
• 5:02
  SPEAKER_02: to be honest.
• 5:02
  SPEAKER_02: I hope so.
• 5:03
  SPEAKER_02: Yeah, more than 100, I think.
• 5:06
  But we also want to talk about some of the interesting findings
• 5:11
  of the survey.
• 5:13
  Most of the questions of that survey are kind of multiple
• 5:17
  choice questions.
• 5:18
  But like many other surveys,
• 5:20
  we have demographic questions at the end of the survey.
• 5:24
  And also we ask teachers to give us additional feedback if they
• 5:29
  are willing to.
• 5:30
  SPEAKER_00: Which is where you get some interesting things.
• 5:33
  SPEAKER_00: And I want to publicly,
• 5:35
  SPEAKER_00: I hope they listen,
• 5:36
  SPEAKER_00: but respondents may or may not,
• 5:39
  anybody who's taking the time to go beyond just answering if
• 5:43
  they've never done it,
• 5:44
  SPEAKER_00: sometimes done it,
• 5:45
  SPEAKER_00: or always done it,
• 5:46
  SPEAKER_00: and things like that,
• 5:48
  SPEAKER_00: you know,
• 5:49
  SPEAKER_00: the Likert scale kind of responses,
• 5:52
  I really would like to thank anybody who takes the time to
• 5:54
  write a few sentences in
• 5:56
  SPEAKER_00: response to let us know something further.
• 6:00
  SPEAKER_02: Exactly.
• 6:01
  SPEAKER_02: Actually,
• 6:02
  SPEAKER_02: we are trying to see how ELA teachers,
• 6:06
  SPEAKER_02: to be more specific,
• 6:07
  are using Gen AI tools to kind of mix culturally responsive
• 6:15
  teaching with the use
• 6:16
  SPEAKER_02: of generative AI in their teaching practices.
• 6:20
  SPEAKER_00: Yeah, and that is the heart of what we're doing.
• 6:23
  SPEAKER_00: So we're interested in examining how AI or generative AI,
• 6:29
  SPEAKER_00: going back to the point I just made,
• 6:31
  generative AI can help you think about your own biases and help
• 6:38
  fix things.
• 6:38
  SPEAKER_00: And I've talked about this before,
• 6:40
  but I think this is a really important thing because we all
• 6:43
  know that generative AI
• 6:44
  SPEAKER_00: can be biased.
• 6:45
  Because it's fed on everything humans have created that has
• 6:49
  found its way to digitize form.
• 6:51
  SPEAKER_00: And therefore, it creates, it absorbs that bias.
• 6:57
  SPEAKER_00: But at the same time,
• 6:58
  SPEAKER_00: by asking the right questions,
• 7:00
  SPEAKER_00: by giving the right prompts,
• 7:01
  it can actually fight against bias and be more culturally
• 7:05
  responsive than you can
• 7:09
  SPEAKER_00: be just on your own.
• 7:11
  SPEAKER_00: So this is really our angle here.
• 7:15
  SPEAKER_00: We're asking teachers about it, but let's be honest.
• 7:19
  SPEAKER_00: Because I like being honest.
• 7:21
  SPEAKER_00: Or at least say the words, let's be honest.
• 7:23
  SPEAKER_00: I'm not sure which one.
• 7:24
  But one of the things we know we're doing when we're sending a
• 7:29
  survey like this is
• 7:30
  SPEAKER_00: we're making people think.
• 7:31
  SPEAKER_02: Yeah.
• 7:32
  SPEAKER_02: Yeah.
• 7:33
  SPEAKER_02: Actually, it was in one of the teachers comments.
• 7:35
  SPEAKER_02: They said, let me see.
• 7:37
  This survey opened my eyes to the many ways that I can be that
• 7:41
  I could be using Gen
• 7:42
  SPEAKER_02: AI to help incorporate more cultural differences into my class readings.
• 7:47
  SPEAKER_02: And I love that.
• 7:49
  SPEAKER_02: I love that comment.
• 7:50
  SPEAKER_00: Yeah,
• 7:50
  SPEAKER_00: so what we're trying to do while we're thinking about these new things,
• 7:55
  SPEAKER_00: we're trying to get feedback from the field,
• 7:57
  SPEAKER_00: but also push some ideas forward.
• 7:59
  SPEAKER_00: And that idea of being culturally responsive is not new.
• 8:03
  SPEAKER_00: Using AI makes it easier.
• 8:07
  SPEAKER_00: And we know teachers don't have enough time.
• 8:08
  We know teachers don't have enough training sometimes to be
• 8:13
  culturally responsive.
• 8:15
  And so this is an opportunity to be able to take a shortcut
• 8:20
  while staying critical.
• 8:23
  SPEAKER_00: What are some other things?
• 8:25
  SPEAKER_02: Let me see.
• 8:26
  SPEAKER_02: For example,
• 8:27
  SPEAKER_02: in one of those questions,
• 8:29
  SPEAKER_02: the question stemmed was,
• 8:30
  SPEAKER_02: I use GNI tools to help me understand my students' cultural backgrounds.
• 8:34
  SPEAKER_01: Yeah, and what did you find out?
• 8:37
  SPEAKER_02: We have a very similar frequency for never, rarely, and sometimes.
• 8:43
  SPEAKER_02: Okay.
• 8:45
  SPEAKER_02: Yeah, so most of the teachers chose these three options.
• 8:50
  SPEAKER_02: Another question that was interesting to me,
• 8:53
  SPEAKER_02: let me see,
• 8:54
  was during teaching I used Gen AI tools to learn about my
• 8:58
  students' cultures and
• 8:59
  SPEAKER_02: languages.
• 9:00
  SPEAKER_02: And again, we see the same patterns.
• 9:02
  SPEAKER_02: Okay.
• 9:03
  you see sometimes early and never chosen by many yeah most
• 9:07
  teachers okay but on the
• 9:09
  other hand i use gen ai tools to develop instruction supporting
• 9:14
  students academic
• 9:15
  achievement that is very different because sometimes and often
• 9:20
  is almost our whole
• 9:21
  sample so when you really think about it teachers are using it
• 9:25
  first and foremost
• 9:27
  For their own purposes and for planning purposes,
• 9:30
  not as much with their students.
• 9:32
  SPEAKER_00: So if you go to the next one,
• 9:34
  SPEAKER_00: right,
• 9:35
  SPEAKER_00: and you say,
• 9:35
  SPEAKER_00: I use Gen AI tools to integrate culturally responsive practices,
• 9:40
  SPEAKER_00: we're getting to sometimes.
• 9:42
  SPEAKER_00: That is the most frequent,
• 9:44
  SPEAKER_00: and it's kind of a nice bell curve distribution,
• 9:47
  but it's different than I use it for my own purposes to make
• 9:51
  sure that I can work
• 9:53
  SPEAKER_00: on student achievement.
• 9:54
  So we can see that that idea of culturally responsive is not
• 9:59
  quite as used.
• 10:02
  Maybe they haven't thought about it,
• 10:04
  or maybe that's not how they want to use it.
• 10:06
  SPEAKER_02: Yeah, yeah.
• 10:09
  SPEAKER_02: We see the same patterns for the next question.
• 10:12
  I use a culturally responsive stance toward Gen AI outputs to
• 10:16
  make sure they are inclusive.
• 10:18
  So as you said,
• 10:19
  we see again a kind of bell curve shape for the responses to
• 10:23
  that question.
• 10:26
  SPEAKER_02: Their responses about their ethical concern was really interesting.
• 10:32
  So for that one,
• 10:33
  we have a kind of a skewed diagram,
• 10:36
  and most of them said always and often.
• 10:38
  SPEAKER_02: When we ask them, I'm concerned about the ethics of using Gen AI tools.
• 10:44
  SPEAKER_02: So ethical concerns are real.
• 10:45
  SPEAKER_02: Yeah.
• 10:46
  SPEAKER_00: Yeah,
• 10:47
  SPEAKER_00: ethical concerns,
• 10:48
  SPEAKER_00: and that actually encourages me,
• 10:50
  SPEAKER_00: because the number is much higher,
• 10:53
  SPEAKER_00: actually,
• 10:54
  SPEAKER_00: than the teachers,
• 10:56
  SPEAKER_00: or it's higher,
• 11:00
  SPEAKER_00: but it also overlaps with teachers who say,
• 11:03
  SPEAKER_00: I use it.
• 11:04
  SPEAKER_00: And that's what we want to see.
• 11:06
  We want to see people who are using AI,
• 11:09
  but are also concerned about the ethics.
• 11:11
  SPEAKER_00: I'm worried about people who say, there's nothing to worry about here.
• 11:14
  SPEAKER_00: There's a lot to worry about here.
• 11:17
  from the energy although that's been getting better but bias
• 11:21
  and is learning really
• 11:23
  happening with my students and are they using it as a shortcut
• 11:26
  and they're not
• 11:27
  learning these are all ethical and and real concerns that i
• 11:33
  want every teacher to
• 11:34
  SPEAKER_00: think about
• 11:35
  SPEAKER_01: Yeah.
• 11:35
  SPEAKER_00: So that to me is encouraging.
• 11:36
  SPEAKER_00: It doesn't mean you should not use it.
• 11:38
  SPEAKER_00: It does mean you should be thoughtful.
• 11:40
  SPEAKER_02: Yeah.
• 11:40
  SPEAKER_02: Yeah.
• 11:41
  SPEAKER_02: Or you are using it, but consciously.
• 11:45
  SPEAKER_02: And we have also another question.
• 11:47
  I encourage my students to use Gen AI tools to explore cultural
• 11:51
  differences.
• 11:52
  And for that one,
• 11:54
  never and rarely have the highest frequency of responses.
• 11:58
  SPEAKER_02: Yeah.
• 11:59
  SPEAKER_00: Yes.
• 12:01
  SPEAKER_00: And I love,
• 12:02
  we talked about this dichotomy between I'm concerned or I'm
• 12:05
  using it a lot for me,
• 12:07
  SPEAKER_00: but not so much for my students.
• 12:09
  SPEAKER_00: There's a response that really, really hit me.
• 12:12
  SPEAKER_00: And this is a response.
• 12:14
  SPEAKER_00: Thank you for responding that way.
• 12:16
  SPEAKER_00: As someone who has been on both sides of education,
• 12:20
  SPEAKER_00: I was a senior in 2020, so about five years out of the university.
• 12:27
  SPEAKER_00: After coming back from COVID, we're playing a different ballgame.
• 12:30
  SPEAKER_00: We need AI to do our jobs, but the children...
• 12:35
  SPEAKER_00: use is to not do their job.
• 12:38
  SPEAKER_00: So the idea is, I need it for my job.
• 12:41
  SPEAKER_00: The kids need to not use it.
• 12:43
  We need to find a way to bring both parties into a mutual
• 12:46
  understanding of what AI
• 12:48
  SPEAKER_00: is and how it can benefit us.
• 12:50
  So it's talking about this exact dichotomy that we're seeing in
• 12:54
  the results.
• 12:55
  SPEAKER_00: I'm using a lot of it.
• 12:56
  I'm not really comfortable with my students using it,
• 12:59
  or maybe their schools doesn't let it.
• 13:01
  SPEAKER_01: Mm-hmm.
• 13:01
  SPEAKER_00: Let us do it.
• 13:02
  SPEAKER_00: And then talking about the productive way we can have a conversation.
• 13:06
  SPEAKER_00: I think this is really the pivotal thing that teachers are telling us.
• 13:12
  And that is we know how to use it to help us with different
• 13:15
  things in education,
• 13:16
  SPEAKER_00: or at least some of them are.
• 13:18
  SPEAKER_00: Yep.
• 13:19
  SPEAKER_00: But how to use it with students is a whole different ballgame.
• 13:22
  SPEAKER_00: And this is where it gets a little bit uncomfortable.
• 13:25
  SPEAKER_00: And rightfully so.
• 13:26
  I think that it needs to be uncomfortable because we actually
• 13:29
  don't know the answer.
• 13:30
  I mean,
• 13:31
  I would like to also always look like the person who has the
• 13:35
  answer.
• 13:36
  SPEAKER_00: I don't think anybody has the answer.
• 13:37
  SPEAKER_01: Yeah.
• 13:39
  SPEAKER_00: All right, so we talked a little bit about our survey.
• 13:42
  SPEAKER_00: Again, if you're interested, just drop us a line in any way you can.
• 13:46
  SPEAKER_00: We would love to have you as part of our participants.
• 13:51
  SPEAKER_00: We want to get a sample as wide as possible.
• 13:53
  SPEAKER_00: If you are teaching English language arts in any way,
• 13:57
  SPEAKER_00: shape,
• 13:57
  SPEAKER_00: or form,
• 13:58
  SPEAKER_00: at any grade,
• 13:58
  SPEAKER_00: we would like to hear from you.
• 14:01
  SPEAKER_00: All right, and thank you for spending some time with us.
• 14:06
  SPEAKER_00: Bye.
• 14:07
  SPEAKER_02: Bye.