Latest / Azadeh and Guy on AI / Fear, Grief, and Agency

Transcript

Fear, Grief, and Agency · Azadeh and Guy on AI

• 0:02
  SPEAKER_02: All right.
• 0:03
  SPEAKER_02: And we're in episode 14 of Azadeh and Guy on AI.
• 0:10
  SPEAKER_02: And today, what are we talking about today?
• 0:12
  SPEAKER_00: Probably we want to introduce ourselves.
• 0:16
  SPEAKER_02: It has been made clear to me that sometimes I forget to introduce myself.
• 0:20
  SPEAKER_02: I'm Guy.
• 0:22
  And I'm a professor at the University of Nebraska-Lincoln in
• 0:26
  innovative learning technologies.
• 0:28
  SPEAKER_00: And I'm Azadeh.
• 0:29
  SPEAKER_00: I'm a third-year PhD student.
• 0:31
  SPEAKER_00: GAI is my PhD advisor, and I study in the same university.
• 0:36
  SPEAKER_02: All right.
• 0:37
  SPEAKER_02: Way to lay all of the responsibility on me.
• 0:42
  SPEAKER_02: So let's start at the beginning.
• 0:44
  We wanted to talk a little bit about a model that we saw and I
• 0:50
  at least liked,
• 0:51
  SPEAKER_02: I don't know if you liked it,
• 0:53
  SPEAKER_02: but I really liked that really helped negotiate for me
• 1:00
  SPEAKER_02: the way people react because we are all AI all the time.
• 1:04
  SPEAKER_02: And even we wake up in the morning.
• 1:06
  SPEAKER_02: Now I'll stop talking about we.
• 1:08
  SPEAKER_02: Even I, I don't need to speak for other people.
• 1:11
  SPEAKER_02: I can barely talk for myself.
• 1:13
  SPEAKER_02: But
• 1:15
  SPEAKER_02: Even I wake up some mornings and I'm filled with dread about AI.
• 1:19
  So there's a lot of emotion that I see in the reaction when I
• 1:25
  start talking about AI to people.
• 1:27
  SPEAKER_02: For example,
• 1:28
  SPEAKER_02: my son,
• 1:29
  SPEAKER_02: the other day I said,
• 1:30
  SPEAKER_02: you know what would be really easy to use for this assignment?
• 1:36
  SPEAKER_02: AI, this is exactly what it was made to do.
• 1:38
  And he just looked at me and started,
• 1:41
  there was anger and dread and a lot of things.
• 1:43
  SPEAKER_02: And he uses AI.
• 1:45
  SPEAKER_02: So it's not the use of AI,
• 1:47
  SPEAKER_02: it's that in the moment,
• 1:49
  you have this emotional response to AI that is somewhat
• 1:54
  different than other
• 1:55
  SPEAKER_02: technologies.
• 1:56
  SPEAKER_02: Although we've always had emotional responses to technologies.
• 2:01
  SPEAKER_00: And most of the time,
• 2:03
  SPEAKER_00: I don't know about your experience,
• 2:04
  SPEAKER_00: but what I experience is that,
• 2:06
  SPEAKER_00: especially with people...
• 2:08
  SPEAKER_00: who are not used to work with AI a lot.
• 2:11
  So when you start speaking about AI,
• 2:14
  there's always,
• 2:15
  they go to the direction of cheating.
• 2:17
  So I have the experience of working with a couple of tutors and
• 2:22
  they are not used
• 2:23
  SPEAKER_00: to use AI that much.
• 2:25
  SPEAKER_00: So when I started speaking about AI,
• 2:28
  SPEAKER_00: first they say,
• 2:29
  SPEAKER_00: oh,
• 2:29
  SPEAKER_00: students are going to use AI to cheat or yeah,
• 2:32
  SPEAKER_00: to help them a lot with writing and everything.
• 2:36
  SPEAKER_02: And that's definitely a worry.
• 2:38
  We're not saying that's not a worry,
• 2:39
  but when we go to that first,
• 2:42
  it's...
• 2:43
  SPEAKER_02: an emotion that we need to work through.
• 2:46
  SPEAKER_02: So Arafa Karimi,
• 2:47
  SPEAKER_02: which I follow,
• 2:49
  SPEAKER_02: I think you do too,
• 2:50
  SPEAKER_02: on LinkedIn,
• 2:52
  SPEAKER_02: has this,
• 2:53
  SPEAKER_02: what she says is it's not a framework.
• 2:56
  SPEAKER_02: It's a way to negotiate really the emotion.
• 2:59
  SPEAKER_02: And she calls it the AI rupture spiral.
• 3:02
  SPEAKER_02: If you're looking, she's from the University of Queensland.
• 3:05
  SPEAKER_02: So we love the work she's doing.
• 3:08
  And she's really active on LinkedIn,
• 3:10
  which is one of those places I get information.
• 3:13
  And she talks about the fact that often the first response is
• 3:18
  fear,
• 3:19
  which is highly relevant.
• 3:20
  SPEAKER_02: And we're not trying to say that these emotions are not relevant.
• 3:24
  I think that the worries,
• 3:27
  are people going to cheat and not do the work,
• 3:29
  are real worries.
• 3:29
  SPEAKER_02: They're not fake worries.
• 3:31
  SPEAKER_02: They're not just a phase you need to get over.
• 3:34
  SPEAKER_02: It's always there.
• 3:36
  SPEAKER_02: But at the same time...
• 3:38
  SPEAKER_02: you have a series of reactions.
• 3:40
  SPEAKER_02: She talks about from fear, we move to grief.
• 3:44
  SPEAKER_02: And this is something I had to learn.
• 3:46
  SPEAKER_02: I don't have that grief for the most part.
• 3:49
  SPEAKER_02: And that is the grief for what is lost.
• 3:52
  When you move to a new technology,
• 3:54
  something is lost in the old way of doing things.
• 3:57
  SPEAKER_00: That's interesting because I didn't experience that kind of grief.
• 4:01
  But I know a lot of people because they love,
• 4:04
  for example,
• 4:04
  the way they used to learn.
• 4:06
  SPEAKER_02: I mean, people talk about, I love the smell of books.
• 4:11
  SPEAKER_02: There's grief with moving to a digital.
• 4:13
  SPEAKER_02: Everybody understands that that's potentially important.
• 4:16
  SPEAKER_02: It doesn't mean that you don't have a little bit of a sense of loss.
• 4:22
  People have talked about the loss when libraries have started
• 4:27
  going away and
• 4:28
  SPEAKER_02: becoming more and more digital.
• 4:29
  SPEAKER_02: Yeah.
• 4:30
  SPEAKER_02: The library right here at the university, there's a Dunkin' Donuts.
• 4:35
  SPEAKER_02: You can buy coffee.
• 4:36
  SPEAKER_02: There are a lot less books available for browsing.
• 4:40
  SPEAKER_02: You can ask for whatever you want.
• 4:42
  SPEAKER_02: It's cold.
• 4:43
  SPEAKER_02: It's cozy, but it's different than it used to be.
• 4:46
  SPEAKER_02: You don't have these carousels and being or corrals.
• 4:51
  SPEAKER_02: I think that's how you say this.
• 4:53
  SPEAKER_02: I don't know.
• 4:55
  And that are kind of deep behind multiple shelves of books
• 5:01
  where nobody can see you.
• 5:02
  SPEAKER_02: I remember in my master's,
• 5:04
  SPEAKER_02: when I was doing my master's,
• 5:05
  I would hide there for days with a pile of books going through
• 5:10
  everything.
• 5:10
  SPEAKER_02: Nobody needs that now.
• 5:11
  SPEAKER_02: I mean, I'm clicking on the computer, I'm getting access to the article.
• 5:15
  SPEAKER_02: Actually, it doesn't exist anymore physically in the library.
• 5:19
  SPEAKER_02: There's a sense of loss there.
• 5:20
  SPEAKER_02: It's grief.
• 5:22
  That's interesting because I think I feel the same loss in
• 5:25
  terms of everything
• 5:26
  SPEAKER_00: being digital,
• 5:27
  SPEAKER_00: but I didn't experience that feeling for AI.
• 5:30
  SPEAKER_02: But if you teach in a specific way and it says all this knowledge,
• 5:35
  all of these ways of being and interacting with students and
• 5:39
  getting their wisdom
• 5:40
  SPEAKER_02: and getting them to think,
• 5:42
  SPEAKER_02: for example,
• 5:43
  SPEAKER_02: let's say using the online vehicles that we have,
• 5:49
  SPEAKER_02: the LMSs,
• 5:50
  SPEAKER_02: to have really productive discussions.
• 5:53
  And what I love about these discussions online is that
• 5:57
  everybody's got the time to think.
• 5:59
  SPEAKER_01: Mm-hmm.
• 6:00
  So it's not that banter in class where some people feel that
• 6:04
  they can talk more.
• 6:06
  SPEAKER_02: Usually,
• 6:07
  SPEAKER_02: even if I have two guys and 28 girls in my class,
• 6:10
  SPEAKER_02: the two guys speak almost as much as the rest of the girls,
• 6:13
  SPEAKER_02: which I dislike horribly.
• 6:17
  SPEAKER_02: So guys always feel like they have something to say.
• 6:19
  SPEAKER_02: Usually they don't.
• 6:20
  SPEAKER_02: I mean, I do love you guys, but...
• 6:23
  SPEAKER_02: Let's be honest here.
• 6:24
  SPEAKER_02: But definitely not as much as everybody else.
• 6:26
  And when there's a group and they're in the group,
• 6:28
  they tend to be the speaker for the group.
• 6:30
  SPEAKER_02: And it's almost natural, but also really, really bad.
• 6:34
  And what I loved about the online discussions is everybody can
• 6:36
  talk when they're ready.
• 6:38
  SPEAKER_02: And if you need more time to plan out or be more thoughtful, you can.
• 6:42
  SPEAKER_02: And that is gone.
• 6:43
  Because I don't know who's giving it to their chat GPT and
• 6:47
  generating something
• 6:48
  SPEAKER_02: that looks very thoughtful.
• 6:50
  SPEAKER_02: Mm-hmm.
• 6:51
  SPEAKER_02: but really isn't.
• 6:53
  And so is there grief there about that opportunity and that
• 6:57
  interaction being lost?
• 6:58
  SPEAKER_02: Absolutely.
• 6:59
  SPEAKER_01: Yes, yes.
• 7:02
  SPEAKER_02: Right?
• 7:04
  And from there,
• 7:04
  she talks about the move to reflection and eventually to
• 7:07
  agency.
• 7:09
  SPEAKER_02: And so she talks about the spiral of emotions.
• 7:12
  SPEAKER_02: And I don't think this is developmental.
• 7:14
  SPEAKER_02: And that's why I like the fact that it's spiral.
• 7:17
  SPEAKER_02: Any morning, any moment, any afternoon, I can be anywhere in that spiral.
• 7:22
  SPEAKER_02: It's really unhealthy.
• 7:23
  SPEAKER_02: But this is happening so fast that I don't think...
• 7:27
  SPEAKER_02: we've had enough time to process.
• 7:31
  SPEAKER_02: So we keep going through it.
• 7:32
  SPEAKER_02: Maybe in a few years, it'll be easier.
• 7:34
  SPEAKER_02: But right now, I think we're going through it.
• 7:37
  And until we explore those emotions and we let other people
• 7:43
  explore those emotions,
• 7:45
  SPEAKER_02: we can't start having a real conversation.
• 7:49
  SPEAKER_02: And what I love about this,
• 7:50
  SPEAKER_02: because one of the critiques,
• 7:52
  SPEAKER_02: for example,
• 7:52
  SPEAKER_02: if you look online and you look at the work she's put out there,
• 7:57
  SPEAKER_02: Some people are saying,
• 7:58
  SPEAKER_02: well,
• 7:59
  I don't believe that this is the way because we have to take
• 8:02
  into account people
• 8:03
  SPEAKER_02: that are taking this in negative ways.
• 8:06
  We talk about the suicides and heightened problematic problems
• 8:11
  in mental health and
• 8:12
  SPEAKER_02: all of these things.
• 8:13
  SPEAKER_02: And I agree that those are things we need to consider.
• 8:17
  SPEAKER_02: But...
• 8:19
  SPEAKER_02: That is also agency.
• 8:20
  It's just agency with negative consequences and maybe
• 8:25
  surrendering agency.
• 8:26
  SPEAKER_02: So it's being lost in the loop.
• 8:28
  SPEAKER_02: But that doesn't take it away.
• 8:31
  That doesn't take away the fact that we have to deal with our
• 8:35
  own emotions.
• 8:36
  SPEAKER_02: How do we feel about that?
• 8:37
  SPEAKER_02: Interrogate our emotions a little bit.
• 8:40
  SPEAKER_02: Allow ourselves to grieve.
• 8:42
  SPEAKER_00: Just acknowledge all of those emotions.
• 8:44
  SPEAKER_02: Yes.
• 8:44
  SPEAKER_02: And to say at any given moment...
• 8:47
  SPEAKER_02: I'm worried.
• 8:48
  SPEAKER_02: I'm worried that my kids will not have a job because of AI.
• 8:51
  SPEAKER_02: That is a real worry that I walk around with every day.
• 8:55
  SPEAKER_02: But I'm completely serious about this, people.
• 9:00
  SPEAKER_02: But at the same time,
• 9:01
  I know they have agency and I know that they're using AI in
• 9:06
  productive ways to
• 9:07
  SPEAKER_02: learn,
• 9:07
  SPEAKER_02: to get better,
• 9:08
  to do their jobs better for as long as they have their jobs
• 9:11
  with the idea that
• 9:12
  SPEAKER_02: maybe...
• 9:14
  SPEAKER_02: Maybe, not for sure.
• 9:15
  SPEAKER_02: As long as they have a way to use AI, they will hold on to jobs.
• 9:22
  SPEAKER_02: That is not guaranteed, but it is a way to mediate reality.
• 9:27
  Again,
• 9:28
  I always go back to maybe in a perfect world,
• 9:30
  we don't have AI,
• 9:31
  but we do have AI.
• 9:32
  SPEAKER_02: It's all around us.
• 9:33
  SPEAKER_02: It's not going away.
• 9:36
  SPEAKER_02: We might be going away before AI.
• 9:38
  Speaking of fear,
• 9:39
  people might be going away before AI goes away,
• 9:43
  but AI does not go away.
• 9:46
  So I love that because it's simple and it is really saying that
• 9:51
  we have to start
• 9:52
  SPEAKER_02: with acknowledging emotions and maybe talking about emotions.
• 9:56
  SPEAKER_02: And I think about our work in the School of Engineering.
• 9:59
  SPEAKER_02: Engineers are not good at talking at their emotions,
• 10:03
  SPEAKER_02: but all of our conversations,
• 10:05
  SPEAKER_02: we're seeing the same things.
• 10:07
  SPEAKER_02: We're seeing the emotions.
• 10:08
  SPEAKER_00: Yes, yes.
• 10:11
  SPEAKER_00: So the other things that you wanted to talk about are two papers.
• 10:16
  SPEAKER_00: Yes.
• 10:17
  SPEAKER_00: And they are also very related to the things that you just explained.
• 10:22
  So the first paper,
• 10:23
  the title of the paper is,
• 10:24
  do you have AI dependency?
• 10:26
  SPEAKER_00: There are also academic self-efficacy,
• 10:28
  SPEAKER_00: academic stress,
• 10:29
  SPEAKER_00: and performance expectations on problematic AI usage behavior.
• 10:35
  SPEAKER_00: There are four authors,
• 10:36
  and I'm going to read all of their names since I don't like to
• 10:40
  say it out.
• 10:41
  SPEAKER_00: So Zhang, Zhou, Zhu, and also Kim.
• 10:44
  SPEAKER_00: And the study has been published in 2024.
• 10:47
  It's been,
• 10:48
  they actually conducted an online survey in August 2023 in
• 10:52
  South Korea.
• 10:55
  The researchers,
• 10:56
  they surveyed 300 university students and all of them had used
• 11:00
  FAT-UBT.
• 11:01
  SPEAKER_00: They wanted to explore factors behind AI dependency.
• 11:06
  SPEAKER_00: And by AI dependency,
• 11:08
  SPEAKER_00: we mean heavy reliance on AI tools for academic work,
• 11:12
  SPEAKER_00: not just occasional assistance.
• 11:15
  And they use a psychological framework and the name of that
• 11:19
  framework is IPACE.
• 11:21
  SPEAKER_00: So it stands for Person, Affects, Cognition and Execution.
• 11:25
  It's a model which is often applied to problematic or addictive
• 11:29
  behaviors to test
• 11:31
  SPEAKER_00: how self-confidence,
• 11:33
  stress and expectations about AI's performance combine to shape
• 11:37
  AI dependency.
• 11:40
  SPEAKER_02: So the key to this is that idea of AI dependency.
• 11:44
  And the advantage and disadvantage of this kind of research is
• 11:47
  that it relies on self-report.
• 11:49
  SPEAKER_02: We do that kind of research.
• 11:51
  SPEAKER_02: So we're not saying it's bad research.
• 11:53
  SPEAKER_02: But we have to acknowledge that this is self-report.
• 11:55
  And that has some implication for whether people are very good
• 12:01
  at talking about
• 12:02
  SPEAKER_02: their AI dependency.
• 12:04
  SPEAKER_00: Mm-hmm.
• 12:05
  SPEAKER_00: So what they found basically is when students experience lower academic
• 12:11
  SPEAKER_00: self-efficacy,
• 12:13
  SPEAKER_00: they actually experience more stress and they highly rely on AI.
• 12:19
  So students who felt less confident about their schoolwork
• 12:22
  tended to feel more
• 12:23
  academic stress and that stress made them more likely to lean
• 12:27
  on creativity.
• 12:29
  The other factor is high performance expectations of AI leads
• 12:34
  to greater dependency.
• 12:35
  Students who believe AI would significantly boost their
• 12:40
  academic performance,
• 12:41
  SPEAKER_00: they're more likely to overuse AI.
• 12:44
  And so I think that it mixes a few things and it pays to think
• 12:49
  about the context of South Korea.
• 12:53
  SPEAKER_02: So South Korea is a place where academic achievement is one of... Yes.
• 12:59
  super important top values and your success at university that
• 13:05
  is articulated
• 13:06
  through grades and placements and all of these things is
• 13:10
  paramount so the amount of
• 13:11
  pressure so when we talk about stress and anxiety we're talking
• 13:15
  about a lot of
• 13:16
  stress and a lot of anxiety i mean i think out there based on
• 13:20
  what i heard from
• 13:22
  SPEAKER_02: yeah
• 13:25
  from some people in our department,
• 13:27
  your academic achievement is a huge part of your identity.
• 13:31
  SPEAKER_02: Yes.
• 13:32
  SPEAKER_02: And that is the context under which we should explore the idea of stress.
• 13:38
  Because I would argue that part of the problem is that when we
• 13:45
  make everything high stakes...
• 13:47
  SPEAKER_02: It's connected to identity.
• 13:49
  It's connected to your future and your ability to earn and your
• 13:52
  ability to tell
• 13:54
  SPEAKER_02: your parents and your parents to tell their friends.
• 13:56
  SPEAKER_02: That chain of pride or lack of pride, right?
• 14:01
  SPEAKER_02: And it's true of,
• 14:04
  SPEAKER_02: it's probably true in South Korea,
• 14:06
  SPEAKER_02: but it's also true of immigrants to the United States,
• 14:11
  SPEAKER_02: of which we are both.
• 14:13
  SPEAKER_02: And so it's like there's this pressure to to perform really,
• 14:16
  SPEAKER_02: really well to justify the immigration,
• 14:18
  SPEAKER_02: but also to maximize and also because of multiple things.
• 14:22
  These are things that are happening in that extra stress is
• 14:26
  what might put you or
• 14:29
  SPEAKER_02: or somebody who is in.
• 14:30
  SPEAKER_02: So the student that is, you know, my kids, for example.
• 14:34
  or in this case,
• 14:35
  university students,
• 14:37
  it might put an extra layer of pressure.
• 14:39
  And then the focus is on achieving the short-term goals because
• 14:42
  it's obvious even
• 14:43
  SPEAKER_02: to students from our interviews,
• 14:45
  SPEAKER_02: it's obvious to them that this is not productive.
• 14:47
  SPEAKER_02: Long-term, somebody will discover that they can't do it.
• 14:50
  But right now there's a grade and we need or want to get an A
• 14:56
  and here's a shortcut to it.
• 14:58
  Even though that we know that once we are put in a room and
• 15:02
  they ask us to do a test,
• 15:05
  SPEAKER_02: We can't use the AI.
• 15:06
  SPEAKER_02: We will not be able to do it.
• 15:07
  SPEAKER_02: And I was talking to my good friend,
• 15:09
  SPEAKER_02: Link Yatso,
• 15:09
  SPEAKER_02: this morning,
• 15:11
  SPEAKER_02: who's a professor in computer science,
• 15:13
  SPEAKER_02: and he was saying,
• 15:13
  SPEAKER_02: this is exactly the problem.
• 15:15
  SPEAKER_02: We can identify that they're using AI, which is fine.
• 15:18
  SPEAKER_02: We're not saying don't use AI,
• 15:20
  but we are seeing that the kids who over-rely on AI tend to
• 15:23
  underperform in the
• 15:24
  SPEAKER_02: test when they can't use AI.
• 15:28
  SPEAKER_02: And we think that some information is foundational.
• 15:31
  SPEAKER_02: And again, even students recognize that.
• 15:32
  SPEAKER_02: But again, when you're under academic stress,
• 15:35
  SPEAKER_02: you want to perform.
• 15:37
  SPEAKER_02: That idea of the high stakes environment is creating,
• 15:45
  SPEAKER_02: people take shortcuts,
• 15:46
  SPEAKER_02: people become defensive,
• 15:47
  SPEAKER_02: people do whatever they need to do.
• 15:49
  SPEAKER_00: Yes, exactly.
• 15:52
  SPEAKER_00: They also talked about,
• 15:53
  actually they reported negative impacts on other reliance on AI
• 15:57
  and you can easily
• 15:59
  SPEAKER_00: predict those negative impacts.
• 16:01
  SPEAKER_00: One of them is
• 16:02
  SPEAKER_00: increased laziness or habit of outsourcing thinking,
• 16:06
  SPEAKER_00: reduced creativity and critical or independent thinking,
• 16:11
  SPEAKER_00: the risk of more misinformation,
• 16:14
  SPEAKER_00: risk of wrong answers,
• 16:16
  SPEAKER_00: and also less information seeking weaker judgment,
• 16:19
  SPEAKER_00: possibility of plagiarism,
• 16:21
  SPEAKER_00: and also loss of problem solving practice.
• 16:25
  And here I think that this is cultural,
• 16:29
  but I would not talk about laziness.
• 16:33
  SPEAKER_02: I would talk about optimizing for reality.
• 16:38
  SPEAKER_02: So if the thing that is most important is getting an A,
• 16:44
  SPEAKER_02: and the A is determined by things that you can use AI to get better at,
• 16:48
  SPEAKER_02: then this is the right approach.
• 16:50
  SPEAKER_02: Again,
• 16:50
  SPEAKER_02: going back to its lack of long-term thinking,
• 16:53
  SPEAKER_02: its short-term thinking,
• 16:54
  SPEAKER_02: but in the short term,
• 16:55
  SPEAKER_02: that works.
• 16:56
  SPEAKER_02: If I can get an A with AI, that's less energy, less anxiety.
• 17:03
  SPEAKER_01: Yep.
• 17:05
  SPEAKER_02: More outcomes.
• 17:06
  SPEAKER_02: So I get the same A, right?
• 17:09
  SPEAKER_02: And this is, I mean, we often talk about Tom Sawyer and the fans, right?
• 17:14
  SPEAKER_02: And he won the,
• 17:16
  SPEAKER_02: if you remember,
• 17:17
  SPEAKER_02: he wins the Bible,
• 17:20
  SPEAKER_02: he wins the Bible award,
• 17:21
  but he can't answer a single question because he got all the
• 17:24
  tickets to win,
• 17:26
  but not in the way that was meant because the emphasis was on
• 17:29
  getting the tickets
• 17:30
  SPEAKER_02: and not on the knowledge,
• 17:32
  SPEAKER_02: right?
• 17:32
  SPEAKER_02: And this is the same thing.
• 17:33
  SPEAKER_01: Yeah.
• 17:35
  SPEAKER_02: Right?
• 17:36
  SPEAKER_02: You can't answer any computer science Bible questions.
• 17:42
  SPEAKER_02: But you got all the tickets.
• 17:44
  So this is also about society changing some of its expectations
• 17:52
  and about how do we
• 17:53
  SPEAKER_02: measure,
• 17:53
  SPEAKER_02: but also how much pressure do we put on individuals on the short term?
• 17:59
  SPEAKER_01: Mm-hmm.
• 18:03
  SPEAKER_02: And the more you put pressure on that,
• 18:05
  SPEAKER_02: the more people,
• 18:06
  SPEAKER_02: again,
• 18:07
  SPEAKER_02: this is the research,
• 18:08
  SPEAKER_02: people with low self-efficacy,
• 18:10
  SPEAKER_02: that's their go-to.
• 18:11
  SPEAKER_02: I don't think I can, what is low self-efficacy?
• 18:13
  SPEAKER_02: It's me saying, I don't think I can solve this on my own.
• 18:18
  SPEAKER_02: So what are my choices?
• 18:19
  SPEAKER_02: Here's something that can do it for me.
• 18:22
  SPEAKER_02: It will relieve all of my stress because it's mediated by stress.
• 18:25
  SPEAKER_02: So I can lower my stress by doing this.
• 18:29
  SPEAKER_02: I mean, it answers itself.
• 18:32
  SPEAKER_02: So it's about thinking about,
• 18:35
  SPEAKER_02: first of all,
• 18:36
  SPEAKER_02: low academic self-efficacy,
• 18:37
  SPEAKER_02: but also how are we sending the signals back to students?
• 18:40
  SPEAKER_02: So there's a lot we can do with it,
• 18:42
  but I think it's fantastic research because it does bring these
• 18:45
  issues to the fore
• 18:46
  SPEAKER_02: and say,
• 18:47
  SPEAKER_02: there's a lot of emotional load that we take and we put on our students.
• 18:56
  SPEAKER_02: All right.
• 18:57
  SPEAKER_02: Do you want to go to the second paper?
• 18:59
  uh sure so the second paper the title of the paper is ai driven
• 19:02
  academic engagement
• 19:04
  in fine arts education the chain mediating rules of
• 19:08
  self-efficacy and achievement
• 19:09
  emotions been published in 2025 and the authors are zhang zhao
• 19:15
  and also lee so what
• 19:17
  SPEAKER_00: they did it's again another survey but they surveyed about almost
• 19:22
  SPEAKER_00: 3,000 fine arts students across 18 universities in China.
• 19:27
  SPEAKER_00: And they asked about their use of AI tools,
• 19:30
  SPEAKER_00: I mean,
• 19:30
  SPEAKER_00: generative AI for art practice or study,
• 19:33
  SPEAKER_00: their sense of self-efficacy,
• 19:35
  SPEAKER_00: and also their emotions about achievements,
• 19:38
  SPEAKER_00: and also their level of academic engagement in their practice.
• 19:41
  So what they found is that AI boosts engagement both directly
• 19:47
  and also indirectly.
• 19:49
  SPEAKER_00: So there is a substantial direct effect.
• 19:51
  AI usage is positively associated with academic engagement in
• 19:55
  arts practice.
• 19:57
  And it means that the students who use AI tools,
• 20:00
  they more tend to report higher engagement.
• 20:04
  SPEAKER_00: And there's also indirect effects through two psychological pathways,
• 20:10
  SPEAKER_00: self-efficacy and also achievement emotions.
• 20:13
  So using AI builds a student confidence in their own artist
• 20:17
  skills and also
• 20:18
  SPEAKER_00: positive emotions tied to creative achievements.
• 20:22
  And what I love about this with the other paper is here they
• 20:26
  defined it a little
• 20:27
  SPEAKER_02: bit differently.
• 20:28
  So in the other paper,
• 20:30
  there was dependency on an AI and that was a negative.
• 20:35
  SPEAKER_00: Unhealthy dependency.
• 20:36
  SPEAKER_00: Yes.
• 20:37
  But it's very clear that their entry point was we want to
• 20:41
  discover who's using too much AI.
• 20:43
  SPEAKER_01: Yeah.
• 20:44
  SPEAKER_02: Here they're saying, before we say if it's good or bad, let's see how...
• 20:50
  SPEAKER_02: it impacts other things that we care about.
• 20:52
  SPEAKER_02: In this case, engagement with art.
• 20:54
  SPEAKER_02: And the answer was,
• 20:56
  SPEAKER_02: it's actually a positive relationship,
• 20:59
  SPEAKER_02: not a negative relationship to an outcome we're interested in.
• 21:02
  SPEAKER_02: So maybe it is about the domain.
• 21:05
  SPEAKER_02: This is art.
• 21:07
  SPEAKER_02: And so that's a very different domain.
• 21:08
  SPEAKER_02: The milieu is a little bit different.
• 21:10
  And so you're really looking at something that is somewhat
• 21:15
  different,
• 21:16
  stated more positively.
• 21:18
  SPEAKER_02: But the emotions and self-efficacy still pay a role.
• 21:22
  SPEAKER_00: They also figured out a kind of a chain mediation model.
• 21:25
  SPEAKER_00: So they said AI use leads to increased self-efficacy,
• 21:29
  SPEAKER_00: which leads to improved achievement emotions.
• 21:32
  SPEAKER_00: And finally, it leads to greater academic engagement.
• 21:36
  SPEAKER_02: Yeah.
• 21:37
  SPEAKER_02: And achievement in motion is an area I chose this paper.
• 21:42
  SPEAKER_02: This is the bottom line.
• 21:43
  SPEAKER_02: I love thinking about achievement emotions.
• 21:47
  SPEAKER_02: Reinhard Peckren started this way of thinking about 25 years ago.
• 21:53
  SPEAKER_02: There's fantastic literature.
• 21:55
  And what it helps is it helps kind of work through that black
• 21:59
  box of self-efficacy
• 22:00
  SPEAKER_02: and say,
• 22:01
  SPEAKER_02: what are the mechanisms that work around that?
• 22:04
  And emotional mechanisms are really good ways to think about
• 22:08
  how self-efficacy
• 22:09
  SPEAKER_02: actually get translated to action.
• 22:12
  SPEAKER_02: or lack of action.
• 22:13
  SPEAKER_02: So it can freeze you up.
• 22:16
  SPEAKER_02: Like, you know, I'm giving up and AI will do everything versus
• 22:20
  I'm doing something else and I'm actually being more persistent
• 22:25
  or more inquisitive
• 22:27
  SPEAKER_02: or all of these things.
• 22:28
  SPEAKER_02: And it's got really interesting ways to think about anxiety,
• 22:33
  SPEAKER_02: which is a short-term predictor and boredom,
• 22:36
  SPEAKER_02: which is an achievement emotion.
• 22:37
  SPEAKER_02: Again,
• 22:37
  SPEAKER_02: it's not the emotions we always think about,
• 22:39
  SPEAKER_02: but boredom,
• 22:40
  which was originally one of the most important emotions
• 22:46
  associated with lack of
• 22:48
  SPEAKER_02: long-term success.
• 22:49
  SPEAKER_02: whereas anxiety was related to lack of short-term success.
• 22:53
  So these are really interesting ways to start exploring the
• 22:58
  mechanisms which we
• 23:01
  SPEAKER_02: process this interaction with AI.
• 23:02
  SPEAKER_00: Yes, exactly.
• 23:06
  SPEAKER_00: Yeah, so this is about the second paper that we went through.
• 23:11
  SPEAKER_00: So what matters for art educators,
• 23:15
  SPEAKER_00: educators and also students,
• 23:17
  SPEAKER_00: AI can be used as a creative enabler,
• 23:21
  SPEAKER_00: not as a crutch.
• 23:23
  SPEAKER_00: So for many fine arts students,
• 23:26
  AI tools might lower the barrier to experimenting to boost
• 23:29
  their confidence and
• 23:31
  SPEAKER_00: also spark engagement.
• 23:33
  SPEAKER_00: And we also learned that emotion matters.
• 23:36
  The largest indirect effect came through emotional boosts like
• 23:40
  enjoyment,
• 23:41
  pride,
• 23:41
  and confidence.
• 23:43
  SPEAKER_00: And the findings can be used for other disciplines.
• 23:47
  SPEAKER_00: So this study is in fine arts and its mediation logic,
• 23:51
  SPEAKER_00: which is AI,
• 23:52
  SPEAKER_00: self-efficacy,
• 23:53
  SPEAKER_00: positive emotions,
• 23:54
  SPEAKER_00: and engagement.
• 23:54
  It may apply to other fields such as writing,
• 23:58
  design,
• 23:59
  performance,
• 24:00
  even STEM fields.
• 24:01
  SPEAKER_02: Yeah.
• 24:02
  SPEAKER_02: And so that's,
• 24:03
  SPEAKER_02: again,
• 24:03
  SPEAKER_02: interesting in comparison to the other study,
• 24:05
  SPEAKER_02: because in the other study,
• 24:07
  SPEAKER_02: AI was what you went to when things weren't working.
• 24:10
  SPEAKER_02: Here, AI is what you go to to gain confidence.
• 24:14
  SPEAKER_02: So instead of self-efficacy leading to dependence on AI,
• 24:19
  SPEAKER_02: because I don't think I can do it,
• 24:20
  SPEAKER_02: this is like me and AI together are capable.
• 24:25
  And which motivates me to keep engaging and becoming more and
• 24:28
  more capable.
• 24:29
  SPEAKER_02: I'm not surrendering my agency, going back to the previous piece.
• 24:36
  SPEAKER_02: I'm not surrendering my agency.
• 24:38
  SPEAKER_02: I'm just collaborating with AI.
• 24:41
  SPEAKER_02: So it's a different way to think about this.
• 24:43
  SPEAKER_02: And it's delightful.
• 24:46
  SPEAKER_00: Yes.
• 24:46
  SPEAKER_00: Yeah.
• 24:48
  SPEAKER_00: Using AI as a thinking partner.
• 24:50
  SPEAKER_02: Yeah.
• 24:50
  SPEAKER_00: Yeah.
• 24:51
  SPEAKER_02: All right.
• 24:52
  SPEAKER_02: So it's been longer than usual.
• 24:54
  SPEAKER_02: It's been 24 moments.
• 24:56
  SPEAKER_00: 24, okay.
• 24:57
  SPEAKER_02: 24 long moments.
• 25:00
  SPEAKER_02: I suggest listening at one and a half time speed.
• 25:03
  SPEAKER_01: Yeah.
• 25:04
  SPEAKER_02: Anyway, we'll see you next time.
• 25:06
  SPEAKER_00: See you.