Audio Podcast

Listen to the podcast embedded below, or open "The Atlas So Far..." in a new tab.

Related Resources

Primers for those who haven't read:

Additional Resources

Clever Hans (the horse that could "do" math)
Historical Mechanical Turk (a mechanism used to trick people)
Present Day Mechanical Turk (crowdsourced AI support)
Marx on the Environment
Review of Marx’s Substructure & Superstructure
BBC4 Radio Marx & Alienation,

Text Transcript

This transcript was generated from automatic speech recognition and lightly edited for clarity. Just an FYI and advance apology for any transcription errors. Matt will get to a better edit soon.

Introduction; Discussion of whether Clever Hans was a Horse

Matt Boyle: What is artificial intelligence? Is it the chat bot you use to generate writing? The media generator you use to create or edit images and video? The algorithm that manages what you see on social media? It is these things, but it's also built from the resources, the earth itself and the labor, the people that walk the earth. From the assumptions we make about how the data that sustains AI is categorize, organize, and streamline or not.

Hi and welcome to this podcast from the Teaching and Learning Center. I'm Matt Boyle, your host. In this recording, I talk with Professor of Philosophy James Lincoln and Dean of Curricular Innovation Denny Fry about Atlas of AI, power politics, and the planetary costs of Artificial intelligence. The book we're reading as part of our book group, James and I discuss what we've learned. Denny joins us from the perspective of someone who hasn't joined the reading yet or has only skimmed it. Reach out to the TLC if you have questions about the book group, would like to join, or would like primers on the text. So everyone's welcome, whether you've read in depth or not. Now, without further ado, here's the podcast.

Hello there, and welcome to this podcast where we're going to discuss the Atlas of AI by Kate Crawford. What did we choose this book for, James?

James Lincoln: Yeah. So we read, uh, previously a book about teaching with AI. It's part of our TLC book club. And what we learned out of that conversation was that there was a gap in people's perceptions about what this technology is and where it sits in the larger constellation of social forces, environmental things, ethical implications, really, how it affects how technology in general affects the way we engage with, perceive and think about ourselves and the world.

So we were really looking for a book that would allow us to take a step back, right? So we were jumping so quickly into what is this technology that I do for us in the classroom? What does this technology allow for us to do in our own jobs, inside education and out? Um, that it really felt like we were moving too quickly to understand what the implications of the technology were, because we didn't have a clear view of what forces have allowed the technology to emerge now.

So Kate Crawford's book of AI has quickly become standard reading and classes on AI ethics, uh, individuals and society, technology kind of classes. And, uh, Crawford is a research professor who is nationally recognized as someone who researches, presents and talks about the social implications of artificial intelligence. I think she's approaching AI from the perspective that it's not merely a term used to describe a part of technology, but to describe the result of an emerging set of forces for industry, power structures and society. All these kinds of things.

So [Denny], I'd be really curious to if you had any thoughts as our non-reader reader. What do you think of the book and what you expect from it? And maybe we can respond to that as we've talked about where we've been in the book so far.

Denny Frey: Yeah, sure. Um, and skim reading it, probably way behind where you all are and processing and thinking through and with Crawford's work. But two things that struck me right off the bat about her book is I got into it. One is the way in which it opens around questions and knowledge and, you know, what is knowledge? And she has this wonderful anecdote about, uh, marvelous hands or I can't remember the actual German. Clever Hans. "Klüger Hans!" [Denny was very insistent on the correct spelling of the German]

Um, who could apparently display amazing mathematical skills.

James: And Hans is a horse.

Denny: Oh, yes, I forgot!

[General Merriment]

Matt: That is an important part.

James: That's a yeah, yeah, Hans could do math. Yeah, but Hans was a horse.

Denny: Yes, Hans was a horse! And so there's this sort of wonderful entry into the book about. Okay, this horse that has knowledge. Human knowledge, right. Uh, and careful discipline, skepticism demonstrates that “No, something else is going on there.” We don't have to go into the details. Get into the book. Read it for yourselves. I'm going to go back and read it more closely for myself, because clearly I left out the key point that Hans was a horse!

James: No, I think we should talk about it, because it's really…you're making a point and it's really informative about the framework of the book. It's like Hans being able to talk…sorry COUNT. This is getting into Mr. Ed territory here. Hans being able to do math and count in response to the prompting the and the people who come in and test them and do all kind of stuff overlooked Hans’s capacity to read the room, to give the response that they were expected to give to people, and what assumptions that makes about what intelligence is, and what that means for how we assess if something is intelligent and what that means when we anthropomorphize or intelligence-ize AI in general. So I don't think we should run fast past that. That's really good.

Denny: But that's what really drew me into the book, because, you know, I did read the teaching with AI, for instance, and there is a general sense out there in the atmosphere that, you know, okay, it's a tool. We've got to use it. We've got to get our minds around it. We have to teach our students to use it. And so the faster we can get our minds around it and become acculturated with it, we will be better at teaching our students to use it.

But there are some fundamental questions like that relationship between intelligence, knowledge, responsiveness, and so that I really appreciated in the opening of the book. Second thing that I notice is that and here's where I'm turning to you too, because you're much deeper into the book. But the way in which she approaches her topic is, um, you know, an atlas, you know, includes lots of maps, right? That's what we tend to think of as, as an atlas. And so as I look at the book, I'm a trained reader, so I skim read through multiple things.

I'm looking at the titles of the chapters and the progression of the chapters. And, you know, she moves in a very small and Marxist way through the material. In other words, let's start with some structure, right? The earth, the environment, and then moves to labor. How do we relate to that environment? And then how do we feel about that environment? It strikes me as a very small and Marxist approach to the material, which I think is a wonderful one. Um, because it is demonstrates that discipline skepticism, I'm going to ask questions, deep, probing questions about what this new technological wonder does and is, but these are, you know, my first impressions just looking at it.

Why an Atlas? Why is that the metaphor?

James: Yeah. And that opens us up to the next question on our list about we want to cover it is why is “The Atlas” the metaphor here. Like what's going on? The way she's looking at lived experience as a structure where you're doing the work, where you're living, how that experience. It's changing. Atlases do something for that. So, like, why, Matt, do you think you remember that conversation?

Matt: I do, we were talking about why did she call it Atlas of AI? And I thought, “well, it's a metaphor that she's using to say she's just mapping out all the areas of AI,” you know. It’s a collection of maps of what AI is and how it changes individuals.

And, um. At first it kind of threw me because it didn't feel like the book was really about artificial intelligence at all, so much as it was about the larger structure [of] the world in general. It's about industrialism. It's about changing the way individuals accomplish things, changing into automated assembly lines.

So I kind of got it later on when she got into classification of data, I think she was setting up a metaphor in the beginning that this is an atlas of all the ways that AI is built upon extractive technologies, although you can also define a technology as a mindset of using certain methods to change labor into assembly lines, for instance, to change production of data at the same.

James: I think you're right on track it because the reality is she describes an Atlas very early on as a tool for cutting the world of joints, right? If you have a map in a classroom that has nations on it from 2024 that set geopolitical domains of power onto the format, right? Like a critical context when you look at a globe. And what the amount of water on that globe is versus a map which distorts certain parts of the world to look bigger than it actually is in proportion to other parts of the globe.

That choice, she's noting about how we represent things in cartography, but also in Atlas creation is not as much about saying, “Look over here. We're dealing with earth issues and issues of land and environmental problems. And over here we're dealing with issues of labor. And over here we're dealing with this in this book.”

It's more about this book is supposed to represent a way of looking at AI that cuts its application as a technology and as a social phenomenon at those joints so that you can see, look, “AI emerges as a technology here. But that being said, in the background, we have to have certain systems in place to support the extraction of materials like lithium to power batteries and other rare earth metals to power computer chips. And so what is going on in that system, and how is that informing what tradeoffs are we making?” Right. Between what we take from the world and what we're putting back into it with AI.

Matt: I also was really struck by what you pointed out earlier on about, um, anthropomorphism. There is a narrative about AI being this kind of co-intelligence you can use as your support system to help you learn. It does inject an aura of inevitability about AI. We all have to get on board with it. What I was taking away is that she was saying, we're skipping what it's doing to the environment. We're skipping what it does the labor forces, we're skipping how it's built upon data classifications, which are, at their core, fundamentally flawed, racist, inaccurate, wrong…

James: Or just weird.

Matt: Yeah, weird. And the part that I thought was distinctive that really stuck with me was that there's this sort of talk of, “okay, but it will get better.” And that's the inevitability paradox of it. And the argument she gets into then classification and data that I thought was distinctive was…Maybe it won't get better. Maybe if you think the only way to fix is to tweak an existing system, that isn't going to fix the fundamental flaw that it’s an extractive one at its core.

Chapter 1 Reflection: Earth

James: Yeah, I mean, it might be valuable because part of the purpose of this podcast is to give a sense for folks who haven't been doing the reading, or haven't had time really to do the reading, to catch up. To see where we've been and what we've learned and and what we've been talking about. We've done at a pretty high level right now. So maybe let's take a second and break down what's happening in each new chapter. So we really covered the introduction and what we've been doing so far.

The chapter one about Earth, right? Makes an observational claim about the impact of technological development on different regions of the world, really through mining. We learned a lot about extractive mining for lithium and other like cobalt was another. And what I really thought was interesting about that is the constant observation she makes in that chapter. I don't know if you remember this, but how the way the technology is put in front of people is in a way that assumes that those trade offs are always worth it.

So what we pull from the land we -- and this comes up again the labor part -- but we pull out of Lithium, we pull out of cobalt, we pull out of strip mining, various areas of our environment. And what we create in terms of fossil fuels is a whole conversation in chapter one about water and how much water it costs to do certain things, this is a very part of the contemporary zeitgeist, about how every search of an AI system costs X liters of water because you have to cool those systems.

It really makes you want to think, like, is using this technology or developing this technology worth the environmental crisis, the community crises that they create? Um, because right now, I can’t remember that paradigmatic town that she describes since it's been a while since I read the chapter, but that town is bustling because of a lithium mine. And when the lithium mines dry, people are going to leave it and it's going to be decimated again.

Denny: Ghost towns; she talks about that. That was one thing I caught in the first chapter that was Ghost Towns.

Matt: Yeah, it's the assumption thing I was talking about there that in a cost benefit analysis, the cost is always going to be outweighed by future benefits that are unclear often. And those unclear future benefits lead to the sort of narrative of more investment. We got to invest more, we got to invest more, we got to invest billions, and we will make trillions. We will hurt the environment, but we'll learn how to fix the environment.

It's this kind of cyclical, circular reasoning. Uh, “harm will be done, but harm will be fixed by the harm that has been done, which will give us the means to fix the harm.”

James: And when we were doing this in our book club Heidi, who was with us at the time, made a really good observation at that. It might be the case of sometimes the extraction is worth it. One of the things that gives me a lot of hope about AI in terms of medicine is its capacity to explore protein sequences and develop new kind of ways of engaging in AI. We were really worried that penicillin and antibiotics were going to go bad, but now we have AI can process through permutations of what that what, uh, appropriate, um, protein antibiotics would look like. And we have we covered more of that in the six months since the creation of AI than we have in the past 25 years, just because we have technology that allows for that kind of permutation.

And maybe that is worth it.

Maybe the amount of value we get as a society come and as a world goes through the roof, when people will have capacity to be healthier. Um, I think we when we were in that meeting, we talked about, well, maybe that's not the best framework to think about, but cost benefit analysis is because we're not separate from our environment. Denny, we've talked about it this -- you and I have on the side -- about what does it mean to be integrated into an environment, into a world with animal creatures. And I'm very curious as to what in your skim reading, I like what was speaking to you from your own background about those issues in this job.

Denny: Well, and I alluded to it in my opening remarks and observations, the Marxist approach, in some ways I can't help but, you know, from my own background as a historian in Central Europe, think about how Karl Marx was his theories about the future. There are great issues with those, but what he did do observe really well was that the mining at that time, coal mining, which was still, uh, continuing apace, that there were fundamental structural changes in the relationships that human beings had with their environment, with the goods around them, with their own labor.

And that's why I sort of see this as a Marxist approach, because I get the sense that she is trying to, to remind us that, that we do have to ask these fundamental questions because the cost benefits analysis, you can't go blindly into it. Maybe there are moments where the short-term negative, uh, is bad. For the long term? Yes. No. Maybe I don't know.

None of us can see the future. And that's one of the things that Karl Marx this big work reminds us is that really, in many ways, it's about sticking in the moment and then using that sort of historical perspective to understand what has gone before so that we can make better decisions. I was about to say rational decisions, but I'll leave it to the philosopher to tell us where rationality is and what rationality is.

James: We did talk about that in our conversation about like, what does it mean to take responsibility for our arrival in this particular moment? And as a geographer, and the parallel there is that we live in an awesome reality, which is the product of our history unfolding. And what does it mean to take responsibility for that?

Reflection on Chapter 2: Labor

James: And I think one of the things we learned in chapter one, which we have into chapter two here, about labor, is that we aren't really thinking about that if we, as Matt pointed out, engage in this kind of extractive, normative process. Right? It's about taking so that we can hurt now to fix the hurt later. It's weird. It feels like it's a little antithetical to what it means to be responsible, at least from an environmental perspective. Yeah. You want to talk a little about the labor chapter?

[Matt fumbles for his primer to remind himself about the labor chapter]

James: So if you haven't read them, we have reference sheets for each chapter.

Denny: Sorry, Matt, while you did do a quick reminder because, again, I apologize for being, uh, late to the party. But I'm also reminded that she speaks to and again, this is, you know, one of Marx's great discoveries is the alienation of ourselves as a result of capitalism, not more than just capitalism, uh, around our fascination with goods and things. And this is a shiny new thing that is fun to play with, that is eating with rose colored glasses and shows great promise for all kinds of things. But we've been down this road before.

James: Yeah, and alienation is a complicated technical term. But for those unfamiliar with it, like how would you describe alienation, Denny?

Denny: Well, uh, for me, from the historical and sort of Marxist perspective, i.e., “I don't have an authentic relationship with the community around me, the environment around me. Now I am more, uh, attached to the goods that are around me, uh, that my sense of authentic identity comes from what I have as opposed to what I am.”

James: Yeah. And from a from a different philosophical tradition that gets described very similarly by saying that you are and you are less yourself because the relations that you're using to articulate a sense of self are hollow or shallow or unconnected with other selves. And so you're alienated, you're a bit alien because you're ostracized from being a member of a community or a group, um, or even in really being in a relationship with others and then therefore a relationship with yourself.

So whatever is blocking the task, that is human existence, desire to seek a relationship with others and themselves and communities, the things that block, that are alienating all of that practice is that track. Is that a fair assessment?

Denny: Yes, absolutely. [Gives a thumbs up]

James: For the record, Denny just gave a thumbs up on a podcast.

Matt: Well, yeah, you know, you have to do visual cues on podcasts from time to time to keep people on their toes.

[General amusement]

Matt: the labor chapter, I guess I engaged with the least of all the chapters because it was still talking about things that I wasn't thinking about as AI…I think which I think is part of the point because it is part of it.

And she talks about the common criticism of Amazon, their tendency towards surveillance and monitoring and the amount of time put into each individual thing. I was put in mind the movie I watch, um, several years ago, uh, Snowpiercer, a dystopian movie about a future train driving its way through a destroyed landscape where people are cogs in a larger machine. And the person running the machine is the sort of figure at the end of the train who placed human beings, children, into little spaces in the machine to take a cog and put a heater and do this.

And he had this strange way of talking and thinking where he would say, “Everybody in their place.” And, um, I think that that kind of is a popular culture representation of that dystopian fear behind which he describes, uh, a mega machine, a pool that uses human labor as the parts of AI and looks at the sort of hypocrisy of saying that the AI is this thing that can do things for you when in reality it's kind of often just gig workers doing little bits for you here, here, here, here. Or what is it, Mechanical Turk, the tool [a crowdsourcing service from Amazon that took its name from a historical fraud case]?

James: Yeah, yeah. We talked about it being very similar to the Hans example. Yeah. So the Mechanical Turk example in the text…

Denny: …is not a horse.

James: Not a horse. You know, it's like a facade. When I think of it, my crane, it's all steampunk and, uh, it's just this, you know the movie Big?

Denny: Yeah.

James: So, the fortune teller machine? Yeah. It's very much in my brain a version of that that is presenting as if it can play chess with and be masters at the process. So people come in and they play chess, which I think has parallels to how AI systems that historically defined intelligence as chess playing. Yeah, IBM's Big Think or whatever it was.

Um, but in the book that this example is designed to say, well, actually, inside that machine is a chess master, like wiping their heads, sweating. Moving in, playing chess against people and winning. Right. It's to maps metaphor earlier. The AI systems that are in place here are utilizing labor in very hidden ways.

Matt: Yeah. And that mechanical turk was able to fool people for decades – and there was also this kind of strange ethnic “other-ization” of it being this kind of something that puts you in mind of the magical being from Big [Zoltar!].

And she brings in the talk about how AI systems are sort of trained by individuals who are doing these little micro tasks to identify whether that's a cat or a dog, and then responding to it to sort of empower the AI and in some cases, just working 24 seven around the clock to impersonate AI is to give the impression of this infallible chatbot speaking to people.

Reflection on Chapters 3 & 4: Data and Classification

James: Yeah, yeah, that's a really good way to transition into that chapter screen for about data classification. Now, at this point in the podcast, the listeners are probably being like, is this book about AI or not? And oh yeah, there's a lot of reasons to feel like the first few chapter of this book are talking about AI, and that's intentional, I think, by the author as like a rhetorical tool to say, we need to start by looking at AI as a social constructed phenomenon.

She even says, when I use the word, I talk about this social phenomenon, right? When I use machine learning or AI uses, she’s using the technical term, about the technology that pumps into your ChatGPT and pops into our systems, because in the data classification chapter, another person might -- I think maybe this was Heidi again saying that this is this is kind of what she wanted to talk about from the beginning, right?

The data chapter talks about, like, you're pointing out that this idea that you're collecting this data, you're accumulating this information. And the assumption is that because the cost is worth the benefit, and the cost is always worth the benefit of being able to collect data as much as we can. The big AI companies are collecting as much data from wherever they can, public domain or otherwise, and even in private contractual domains without much regulation, because that's a good. Like they see that as an imperative to make the world calculable.

I have enough data points so that the world can be perfectly simulated in an AI system, so that we can have perfect access to the knowledge of the world. So the data part's really important because it talks about what privacy violations occur. This is the real meat and potatoes of what AI ethics gets into that I think we all want to talk about, but it really doesn't hit as hard if you don't have a chapter about what the extractive mentality is and a chapter on how that extractive mentality shifts into dehumanizing people into cogs, and then saying, our digital fingerprints are just like those labor fingerprints, and that it doesn't really matter what my privacy is related to something else, if it's for the greater good of extracting, for the sake of training, a system to make deepfakes, to undermine political institutions and then the classification of that data is even more problematic because that classification, as according to earlier, Matt, is like someone sitting in the room clicking on things that are cats, clicking on things that are dogs, and being given categories to say this stuff belongs in this bucket, and this stuff belongs in this bucket.

And that's just reintroducing social assumptions about what we think things should be good/bad, black/white, male/female, all these things.

Matt: She makes a clear comment about the lack of non-binary categories in these, but also the bizarre categories, like there's a picture of a kid from school and real visual markers about who he is or anything, just a kid from school.

And his category is “loser,” which…Beyond being problematic is also confusing.

James: Which is like every 1980s film. It’s really baffling that’s the category, and it's really problematic. I think that last thing I'll say is that it falls into the labor conversation again as being a required conversation before we get here, because we assume that we just give data to we're saying, and there's no parameters or corrective stuff, like we're hiding our classification processes by putting them in a system and forgetting about that system.

But we're also hiding human beings from doing the work of classifying and training systems through human feedback systems that get these, yeah, that get this technology to “hit the target”

Denny: I was just going to remark on the tension that's still is is in place in our contemporary society between, you know, the sort of modern science. And by modern science, I mean 20th century science, late 19th, 20th century science that believed in verifiable, objective truths and that those truths tended to be across a, you know, a binary, if you will, um, right or wrong.

And, you know, late 20th century philosophers, social scientists studying how science actually works. And there's a whole raft of folks, philosophers of science do this all the time.

James: Thomas Kuhn.

Denny: Right. You know, arguing that objectivity is impossible because of human bias.

So it's much more about recognizing the bias, recognizing the subjectivity that that we are prone to.

And so that classification is a process that is embedded with all kinds of problems. And if we aren't careful, uh, we reify those very same problems of creating binaries for the world, when in reality the world is much more complicated and nuanced than male, female. Black/White. Good/Bad. Loser/Winner.

You know, there are lots of winners who feel like losers at the end of the match, the game, whatever the case might be. These are definitions that are oftentimes fungible and that, you know, and I'm not yet as deep into those two chapters, but, I'm excited by what she's, she's trying to, to convey to us is that, you know, we are still there's still some tension here.

You know, I always think of the surgeon's motto that, you know, you have to harm to do good, and you've got to cut to the good.

Matt: [thoughtful] Yeah.

Denny: And, well, you know, not all doctors believe that.

Matt: [bemused] Yeah.

Denny: That's a that's a Western idea that develops in the 19th century as surgeons really start to, to professionalize. There are plenty of medical traditions that don't rely on doing harm first to make good. Um, and it's recovering that and reminding ourselves that that, you know, we have inherited these systems and especially around technology and modern science, that that tend to say, okay, let's do a cost benefit analysis, uh, and let's move forward with this. Let's gather the data because the data is always objective.

And that's false. Yeah. It’s always biased.

James: Yeah. And I think…well, biased has a negative implication to it here in that implicit biases and things are harmful. But bias in the way I think you're wielding it, Denny -- and help me if I'm wrong -- is an indication of a standpoint. So and from that standpoint, epistemologies and black feminist epistemology, like the idea of a standpoint…like, intersectional kind of theories of identity and intersectionality, and the analytical tool for looking at systems of oppression, like bias plays a role in articulating what your standpoint is.

But it doesn't mean that values or ideas or harms are all relative, right? Um, when I do ethics classes, I always try to emphasize to students that there are a plurality of beautiful and wonderful, awesome ways to live. And those can all be good. But there are also some ways that are living that are not good. And it's a it's a harsh example, but one of the things I say in the class is like murder is a bad, it's a bad. Socially, it's a bad regardless of where you are, how we identify it.

And it might have some social and cultural nuances to it, but we still identify as a bad across disciplines. And if you're living a life where you're just murdering, that's not a good life, right? But we can say that at the same time, respect standpoints in that in a way of retrieval that you described. Is that fair?

Denny: Absolutely. [doesn’t thumbs up, but you can tell he wanted to]

Looking Ahead

Matt: I did skim ahead reading the first five pages of each chapter, and I was struck by how a lot of this is starting to be looked at in the chapters on affect and the chapters on state. And it's a little bit like someone rented Minority Report in Spielberg's movie with Tom cruise and decided that they were going to, say, “Let's predict a person's future, let's create classifications for them, binary ones that are accurate and objectively true -- to what you said [Denny]. That may not be possible to make -- and it limits what the “good lives” can be, because it tries to take away the language to articulate them in many ways. It says there are only this many emotions and they are the same across cultures.

I’m put in mind of a Ring Lardner play. The stage direction reads something like, whomever walks on stage, “She looks as if she has eaten waffles.” Which. What the Hell are you talking about? How do you play that? How do you be a person who has eaten waffles?

Like the relational classifications that don't make sense to classify data and speak to whether or not there is an underlying problematic nature to the classification that can never be gotten away from. And that is going to create futures for people and seeks to create futures for people based upon erroneous perceptions of things. Facial recognition to tell you who you're going to be in the future and who you can't be. Well, that's kind of a dark perspective to leave on.

James: Yeah, way to bring us down into the muck.

Matt: Well, you know, I'm the one with dystopia today.

[Further amusement]

Denny: I'm struck by -- and again, I don't have the depth of reading you both have – but since we are relying on some pop cultural references, you know, I also, uh, in, In Thinking about Elysium, the film with Matt Damon and Jodie Foster, where, you know, it's a dystopian future where, you know, there are an elite society of humans living off Earth because we have rendered earth, uh, nearly inhospitable.

And then there are orphans, desperate folks living on what is our home. Um, but their lives are short and brutal. Nasty and, uh, and the lives and in in space in this almost certainly built by spaceships are wonderful.

Matt: I'll just say the last chapter of the book is space.

Denny: [laughs] Oh no!

Matt: Yeah, and it uses Bezos's love of space in a similar way.

Big Takeaways with James and Matt

James: Maybe to wrap this up, we can talk about what's our big takeaway from the conversation, or from what we've read so far? And I have not skipped ahead. I'm only up through the classification chapter, so I'll be reading chapter five for next week.

I think the big thing I'm taking away from this so far is how valuable it is to try to get to using your word to retrieve or remember the forgotten things, uh, that have played a part in any creation, whether it's the extractive nature, what does the environment or the town or the people are in those towns, they extractive labor, uh, the people who are doing the work behind the scenes to build these things.

And it was just a report, uh, in the spring, about AI developers quitting because they're really concerned that they're not putting guardrails on how they're developing this technology, and they're just being asked to work more and more and more for the sake of developing the next model for more efficient outcomes.

And the hidden social structures that we just assume are natural, get erased from our memory by being put into our social systems, whether it's our school systems or whatever, in the same way that data and classification are being used to race our memory of really problematic social structures like racism, sexism, bias, not binary thinking about gender, those kind of things, by letting the system adopt patterns that reinforce and perpetuate that state.

So I'm really taking away from this so far an idea of what it means to remember.

Matt: I agree, I was put in mind of the study about, um, resumes that are perceived as not as good because they have the word “woman” in them, or they have a name that sounds vaguely ethnic, and using a system, an algorithm to sort of sift through them -- just because there's so many -- tends to reify categories and structures and also add an additional element of “it's not my fault. It's the tool.”

“It's a neutral tool.” And I guess my takeaway is if you set up the tool or use it, what is your responsibility to redefining that neutrality? How responsible are you for the use of a thing that you might otherwise say it's accountable, not me.

James: Definitely implications for our are plagiarism policies.

Matt: Yes.

Denny: right.

Matt: So you are welcome to join us now moving forward. We've got four chapters left, and then we're probably going to organize some kind of event where individuals can come in and discuss the book as a whole.

We meet Mondays at noon in the Teaching and Learning Center. You’re welcome to join us if you can. If not, you're welcome to join us in the asynchronous forums and talk online. Thanks for joining us, everybody. And Denny and James, thanks.

Denny: Thank you.

James: Thanks for hosting.

Matt: Yeah of course. Bye bye.

Creative Commons music by Jason Shaw from https://www.audionautix.com