top of page
Search

The Good Robot Hot Take: Most AI Scientists on Screen Suck

Welcome to our second episode of the Good Robot Hot Takes, where every week Kerry and Eleanor give you their spicy opinions about top issues in tech. This week we talk about science fiction films, why we love Aliens and Sigourney Weaver, how female AI scientists and professionals are represented on screen, how this contributes to the unequal gender dynamics of the AI industry, why Iron Man's Tony Stark sucks, and why he and Ex Machina's Nathan Bateman aren’t just bad apples but an epidemic of conceited AI scientists on screen.


Reading List


Cave, S. J., Dihal, K., Drage, E. and McInerney, K. (2023) Who makes AI? Gender and portrayals of AI scientists in popular film, 1920–2020 https://journals-sagepub-com.ezp.lib.cam.ac.uk/doi/10.1177/09636625231153985





Cave, S. and Dihal, K. (2023) Imagining AI: How the World Sees Intelligent Machines: https://global.oup.com/academic/product/imagining-ai-9780192865366?cc=gb&lang=en&



Cave, S., Dihal, K. , Drage, E. and McInerney, K. (2023) 'Shuri in a Sea of Dudes: The Cultural Construction of the AI Engineer' in Feminist AI, https://global.oup.com/academic/product/feminist-ai-9780192889898?q=feminist%20AI&lang=en&cc=gb






Eleanor Drage, 'The Planetary Humanism of European Women's Science Fiction: An Experience of the Impossible' https://www.taylorfrancis.com/books/mono/10.4324/9781003398103/planetary-humanism-european-women-science-fiction-eleanor-drage


Kerry McInerney on BBC Radio 3 Free Thinking, 'Yellowface, AI and Asian stereotypes', https://www.bbc.co.uk/programmes/m001md9d


Kerry McInerney on BBC Radio 3 Free Thinking, 'Introducing New Generation Thinkers 2023' https://www.bbc.co.uk/programmes/m001khj3


TRANSCRIPT:


ELEANOR DRAGE:

Welcome to our second episode of the Good Robot Hot Takes. Every two weeks Kerry and I will be giving our hot take on some of the biggest issues in tech. This week we talk about science fiction films, why we love Aliens and Sigourney Weaver, how female AI scientists and professionals are represented on screen, how this contributes to the unequal gender dynamics of the AI industry, why Tony Stark sucks, and why he and Nathan Bateman aren’t just bad apples but part of an epidemic of conceited AI scientists on screen. We hope you enjoy the show.


KERRY MCINERNEY:

 Hi, and welcome back to our second ever Good Robot Hot Take where Eleanor bring you our spicy opinions about whatever happens to be on our minds. And so this week we are talking about films about AI and specifically Hollywood films. So Eleanor, when is the last time you watched a film about AI?


ELEANOR DRAGE:

I'm so excited for the spicy takes! this is a good question. We're watching them at the moment for a paper that we're gonna talk about, but I'd probably rather say, why did I even start enjoying science fiction films in the first place? And I was thinking about this earlier today, and I remembered the planetarium, this amazing place on Baker Street where you go when you are little or as a school trip and you sit in this dome like building and you look up at this massive screen with stars and it explains about the solar system and how it works.


And I found it truly magical, and that's what I'm looking for in a science fiction film. It's really the technology that is exciting, that is a different way of consuming a film. You are enthralled, not just by the storyline, but by CGI, by whatever is the most innovative technology that is supporting this kind of film today.


So I know that I am fundamentally a theorist of gender. But I usually put that hat to one side when I go to the films and I want a total suspension of disbelief and just to go all in for something.


So the last thing that I watched was Aliens. And I love Sigourney Weaver. I'm just such a massive fan. I love that she stares out the Alien. You have that like woman on woman moment, when they're just looking at each other like, yeah, we're both really powerful, but one of us has to die. It's like Harry Potter and Voldemort, but the original version and much better.


KERRY MCINERNEY:

Yeah, so next week's Hot Take is actually just Eleanor and I looking at each other and deciding which one of us has to die. Yes, I watched Aliens for the first time a week or two ago, again for our research. It was so good. I've never seen any of the Aliens series and I was like, Ugh. Like I don't really, I'm not a big film watcher as a person, but this, I was like, it was too scary for me cause I'm also a wimp, but genuinely I was like, this is such a good film.


ELEANOR DRAGE:

Yeah, there's something great about the chest bursters that come out of the men as well, and I used to describe it as inverse rape, and I still maintain that it's fundamentally a film about colonization. It's the humans that want to go and take over that planet, they wanna make it habitable.


They send over families to terraform it. And you are positioned to make you feel bad for the human families, right? Cause they bang on about the children that are there. You have this little girl. But really the aliens were doing fine. Like, okay, they're gross.


But we could have just left them alone and let them do their own thing. So you have that tension like, who are we supposed to feel bad for? And that's what's amazing too about Octavia Butler and her Xenogenesis series. You have these really grotesque descriptions of alien- human intercourse, where the aliens don't let humans copulate with each other without alien intervention.


But really at the end of the day, the aliens kept the humans alive. So I want a science fiction film where you don't really know who to vouch for, and part of you is a bit like, maybe I am an alien.


What was the last film you saw?


KERRY MCINERNEY:

Oh, well, unfortunately the last film that I saw that had AI in it was The Amazing Spider-Man 2, which was not a rip roaring ride. It was a very, very long film with people who clearly looked like in their thirties playing 18 year olds, which just the older I get, the more and more I'm like, I cannot suspend my disbelief enough to believe that these people, it's easier for me to believe that Andrew Garfield is Spider-Man than it is for me to believe that he's an 18 year old. it was fine, it was just like a very classic like AI, female service assistant, the kind of trope that we see time and time again and stuff like Spike Jonze's Her, and you know, the Ironman series, with Friday. So yeah, I think it was not particularly riveting, and again I mentioned I'm not a big film watcher, but at the same time I think it's really important we talk about films because they have an incredible power when it comes to shaping our cultural understanding of what AI is, and also in quite directly driving and interacting with forms of tech development.


I know that's something you are really interested in, Eleanor, this kind of feedback loop between Silicon Valley and the kinds of films that made in Hollywood.


And also on the Andrew Garfield note, I think what's really annoying is that the narrative of this film is that men can't do what they're supposed to do, like the chores, the duties, the obligations to their relationship to the people closest to them. Like he can't turn up to Emma Stone's graduation because he's too busy saving the world, you know?


And it's like, okay, most men aren't saving the world, but it's just a justification of not fulfilling their obligations. That's an aside.


So the feedback loop I think is really interesting. I found all these things online about Elon Musk's favourite books and the top 10 favourite books are all kind of the same.


They're all Asimov, same kinds of writers of science fiction. I think they're all male, could be wrong, but there's loads of science fiction out there and it's really sad that he is reading a similar genre of book from a similar period, and there's loads of different kinds of works from across the globe from different kinds of writers that remain unread.


The problem with that is that those films don't get transformed into products. So, Josh Wolfe, who's another investor in AI, also reads similar books to, to Musk. He reads a lot of Asimov and then he invests in a lot of companies that have names from Asimov's works, that bring to life kinds of robotics and kinds of innovations, under the same kinds of principles, those three laws of robotics and this means that the things that we have in our daily life, the technologies that we use, all come from the same kind of imagination, this quite masculinist mode of imagining what is possible.


KERRY MCINERNEY:

Absolutely. And I think the interplay between fact and fiction here is just extremely weird, and it's something that people will often say, oh, well, why study science fiction? You know, these are just stories. But they have a real tangible impact, as you've said it, on what kinds of projects get funded and also what kinds of frameworks people have for thinking about technology all the way through to like, who gets involved in these different kinds of AI ethics projects.


One of my favorite kind of weird overlap facts is that Morgan Freeman, who plays a character in the very bad 2014 AI film Transcendence is now on like the board of Advisors or something of the Future of Life Institute, which thinks a lot about AI risk, which is something we talked about in our last Hot Take on these big open letters about pausing AI.


So definitely check out that episode. But isn't that just so bizarre that a Hollywood actor who I assume is definitely not also an AI ethicist on the side, I don't think Morgan Freeman needs to be doing that now sits on an advisory board.


Robert Downey Jr. Deliberately based Iron Man's character on Elon Musk, right? And there was this very deliberate interplay between the two of them to the extent that Elon Musk then appears in Iron Man 2 as a kind of some kind of background actor.


I can't remember the exact role. But it's really interesting to think about this character who's very much portrayed as a hero and like the beating heart of the Marvel cinematic universe, very, very culturally influential, possibly, almost entirely responsible for reinvigorating the career of Robert Downey Jr, was based on a figure who we would possibly say has had some very public falls from grace recently, but maybe as an AI figure or tech figure that we don't necessarily want our quote unquote heroes emulating.


ELEANOR DRAGE:

Tony Stark is this classic genius, solitary figure that works alone. He's not very nice. You know, he's kind of a dick, a bit self obsessed, and we don't want to encourage people to think that they have to be like that in order to be a genius or an innovator, you can be nice and work in a team. So maybe that leads us well onto our own findings from our research paper on this interaction between film and reality. So, Kerry, describe it to us. What did we do? What we find?


KERRY MCINERNEY:

So the theme of today's episode is, as you can tell from our title, most films about AI are boring. And I just wanna clarify, the research study we've done was not 'are films about AI boring?'. So obviously that would be a very subjective study. But what we did instead is we decided to try and look at how AI scientists are represented on screen and these very popular or big budget or influential.


Films and if you're really deeply interested in the methods and how we decided which films fit in that corpus, you can go read our paper. It's all available on public understanding of science and we'll have the article linked in the show notes.


We know that the AI industry is a place that has been very, very inhospitable to certain people that white men are way overrepresented, women and people of color are grossly underrepresented. And we think one driving factor behind this is what we call the cultural construction of the AI engineer in this paper, which was co-authored with Dr. Stephen Cave and Dr. Kanta Dihal. And so we thought, okay, has anyone actually comprehensively mapped how AI scientists are portrayed on screen, and particularly how their representations correspond with gender.


And I think we already had quite negative opinions in mind about what we thought we'd find, but we really were quite blown away to find that of 142 films that featured AI of which 86 showed an AI researcher, or more than one AI researcher. We found that only nine out of these 116 scientists were women.


So that's 8% and also that none of these films were solely directed by a woman. So you see this kind of environment being portrayed on screen, which is heavily exclusive to the extent that women scientists were even more poorly reflected on screen than they are in life because the number of women in AI is very low, but I think globally, about 22% of AI professionals are women. So there was a stark gap between what the reality of the industry is and what people are seeing on screen.


So it's not just that cinema mirror's reality, but is actually substantially worse, which tells us that, as you were saying, in the popular imagination, there's something about AI , what it stands for, what it looks like, what it conjures within us that makes us feel that men should be doing it.


KERRY MCINERNEY:

Exactly. And also important to note that AI and computer science did significantly worse in other kinds of stem. So science, technology, engineering, mathematics, figures, right? So again, female scientists are just very badly represented in general, but there was a really big gap here, even further than that when it came to computer sciences.


And it hasn't always been this way. Right. That's what Mar Hicks was saying when we talked to them a couple of weeks ago. Do you wanna just tell everyone?


KERRY MCINERNEY:

Of course. Mar Hicks is an amazing historian of technology and we have a podcast episode with them which we're very thrilled about. Mar Hicks has traced in a book called Programmed Inequality, the history of computing in the uk, specifically how computing was originally understood as feminine work as a secretarial profession that wasn't very well remunerated, wasn't very well respected, and how in the eighties it transformed into this very masculine profession as it became more societally valued, more seen as a technical field, and started to pay a lot more. And Mar actually argues that this is one of the major reasons why Britain stopped being a world leader in computing this gender dynamic change.


And again, it's also important to note that not only did computing used to be a much more heavily female, uh, field, there's also a lot of countries where women are much better represented in computer science. The US and the UK, I think are just particularly bad examples when it comes to representation.


Yeah. Do you remember we talked to some women from Indaba X, so Indaba X helps build capacity in machine learning in Africa, and we were talking about discrimination in STEM subjects in the uk and the women we were talking to said, well, Indaba X's organizing committee is mostly female. We were like, okay. So really it's a problem that looks different across the globe.


KERRY MCINERNEY:

And that's why I think it's particularly worrying though, that these very culturally dominant Hollywood films, because. We looked also at these like global lists of highest grossing films, that the films that, as Eleanor mentioned earlier, are really culturally prominent. The kinds of sci-fi stories that get told on the big screen are very much coming from the West, and they're very much reproducing these harmful American and British stereotypes about who counts as an AI scientist.


And we got to see firsthand the effect that this has because Stephen Cave, who's the Director for the Center of the Future Intelligence, where we work, he was watching these films with his three daughters. And when you are showing them Tony Stark- like AI creators and they're not seeing themselves represented back in these figures, you can see what it means to them. You can see in their reactions, in their projections of their own ideas onto these films and vice versa, that it really does impact young people and you have these really heavily militarized contexts in films. What's the best example of a really militarized science fiction film with AI ?


KERRY MCINERNEY:

Oh, I actually do have an interesting example here, but it does come back to the Marvel Cinematic universe again. This is my main body of knowledge for this paper, it turns out, which is that different films in the MCU actually worked with the Pentagon, right, to be able to use some of the military materials, or there's this like really, really strong, interrelationship between Marvel and Disney and the military industrial complex in the US which is wild because these are some of the most high budget, widely watched films in the world.


Yeah, so when we're thinking about gender, we're not just trying to do a headcount of men and women. We're thinking about the masculinization of AI and also the feminization of certain tropes around self-sacrificial women that have to die in order for humanity to be saved, which is what happens in Transcendence, this godawful film with Johnny Depp. That could be its own episode of us as bitching about Transcendence. And you have Rebecca Hall, who is a really intelligent -seeming person and it doesn't really make sense on screen for her to be handing Johnny Depp his clothes and introducing him at seminars.


You really feel like it should be the other way around. So we're really attentive to what counts as a masculinized and a feminized film, rather than just doing what you'd expect gender scholars to do, which is do like a duck duck goose with men and women.


KERRY MCINERNEY:

Yeah, exactly. So we definitely did the duck, duck goose, we counted because we were like, look, it's imperfect, but we do think it's important to get some kind of quantifiable sense of how these scientists being portrayed. But we also looked at a number of different kinds of gender tropes that arose from the film.


So one of them, as Eleanor's mentioned, was this idea of the genius. And so that was about a third of our creators were portrayed as Intellectually extraordinarily brilliant. And that is a massive problem because as a number of scholars have pointed out, and I'm gonna link their papers again in the show notes, there's an idea called the brilliance bias, which suggests that women tend to be discouraged from entering careers where it's perceived that you need to be a genius or have really unusually high intellect to succeed in it.


So this tends to include fields like philosophy or mathematics, and yes, also computer science. And this is not because, people tend to think that women as a whole are less intelligent than men, although I'm sure we've all met some people like this. But more that there tend to be this broader stereotype that extraordinary brilliance is more associated with men.


And we could maybe attribute this to a very biased historical record where people like Einstein and Newton all tend to be men because women were systemically blocked from education among other things. And so we think that it really matters that we are portraying AI scientists as these individual geniuses on screen, these child prodigies people like Tony Stark, who can invent an element in one night, like really wild things. And look, I'm not saying that we should have to watch films, which very slowly watch someone work in the lab for 40 years. It's more just that this is so unrepresentative of how these products are made, the teamwork that's required, and just the slowness and the frustration of AI development.


It really makes it seem like magic, and that is such a harmful myth to be propagating.


Yeah, and even the first research paper that I could find published by OpenAI was co-authored by something like 20 to 30 engineers and other team members. So it's a massively collaborative endeavor. So you can see that what we're trying to do is two things. We're trying to do some counting because people like statistics, but then we also do this more qualitative analysis where we look for examples, tropes, that kind of thing, so we can make our argument in many different ways. We were asked a lot, not just about the creators, but about the bots themselves. How is AI gendered? What does AI look like in a lot of these films? I was talking to a guy at the BBC who was saying how many of the these AI humanoids are women and how many are sexualized? And immediately quite a few pop to mind.


Go on, Kerry-


KERRY MCINERNEY:

definitely Ava from Ex Machina that to me is just one of the epitome of this very white, hypersexualized, hyper-feminine kind of android.


Yeah, for me it's Ghost In the Shell, Scarlett Johansson's naked bodysuit. And actually now is a good segue into thinking about techno-orientalism. So maybe in relation to Ghost in the Shell, which probably quite a few people have seen. Can you explain to us what techno-orientalism is and how that relates to Scarlett and her nudity?


KERRY MCINERNEY:

Yeah, of course. And so something I think about a lot, right, is how in a lot of these science fiction films, they're preoccupied with this question of what does it mean to be human in a technological age, and what happens if you can no longer separate out technology from the machine? But the way that they often communicate this idea is through the use of Asia as a kind of background, or asianness as a coded metaphor for an age where humanity and technology are no longer identifiable or separable from one another.


And so Asia comes to represent this dystopian techno forward future. And I argue that this reflects a really, really long history of how Asian people have been racialized in the West as being somehow machinic technological, and this can be linked to this long history of labor extraction under capitalism, where the mechanized body of the 'coolie', quote unquote was seen as being this sort of insensate, ever- suffering, insensitive to pain body that could labor unendingly under capitalism. And this, I think, is really prominent, these ideas and things like Blade Runner, the whole kind of cyberpunk genre, which emerged in the 1980s in response to the so-called Japan panic or the specter of a rising Japan. But we can also see it very much in films like Ex Machina, where we have, um, Kyoko for example, the Asian gynoid who, we don't actually know in the film till halfway through that she's a robot. Whereas Ava has shown from the beginning as robotic. Kyoko, we just assume, apparently the audience does that because she's quiet and she's servile that she's just playing the role of an Asian woman and the really harmful gendered and racial stereotypes associated with that position.


So it comes as a shock, then halfway through the, film when she takes off part of her face and it's revealed that actually like Ava, she is also a robot. So one way that Kanta and Stephen put it, our co-authors is that there's multiple kinds of Turing tests going on here, multiple tests of what it means to be human, but they're very much bound up in these gendered and racial norms.


So beautifully put, and for everyone, this is cutting edge theory. Kerry is writing a book about this. She talks a lot on the BBC about it as a new generation thinker. So listen to all her free thinking episodes if you're interested to find out more.


KERRY MCINERNEY:

Thanks. Oh, I just realized I never actually talked about Scarlet Johansson, which was like your original thought. Yeah, I mean, firstly, this movie is a mess. It takes all the things that are really interesting about the original Ghost in the Shell, and just like Hollywoodifizes and turns it into this like liberal narrative of Scarlett Johansson being empowered and like that doesn't even begin to deal with all the complexities of politics that surrounded Scarlett Johansson's casting as Major Motoko Kusanagi or The Major as she got rebranded. It did spawn though less academically, so many of my favorite memes of Scarlett Johansson as a member of the Asian community. So if you really love mean memes, I would recommend all of this material.


I have not seen those memes, but I will be tentatively looking at them after this. I think you'll probably send like 15, so I won't even have to...


KERRY MCINERNEY:

definitely. I think the question of Scarlett Johansson is really interesting though because I think something that you said to me that I found really striking was how much more likely it is for women to see themselves on screen as a sex bot than as a scientist.


Yeah, or married to a genius. I think like even in real life, you see, um, oh my God, my semi knowledge of supermodels, but not really quite, what's a face that's married to the Snapchat guy?


KERRY MCINERNEY:

Oh, is that Miranda Kerr?


The supermodels are marrying the geniuses. They are the new genre of sportsmen, right. And it's fascinating to watch, but at the same time, they're getting these ideas from these movies where, there is the wonderful, beautiful, very intelligent Rebecca Hall, married to Johnny Depp. I wanna see a film where it's The Bachelorette, but she's an AI scientist.


KERRY MCINERNEY:

Yeah, I wanna see a film without Johnny Depp first. Just in general, not a fan of Johnny Depp, but also, it is interesting that you frame it as this, 'why do they all end up married to these scientists?' Cause it's also something that we identified in the films, right? And this is a trope we called Womb Envy.


The idea that a lot of these AI scientists are not only portrayed as men, but they're also portrayed as creating technology in their own image. And there's this like God complex that plays into it. And I think that's so strong in Ex Machina, like you see Nathan, who's very unambiguously portrayed as the bad guy, but he like talks about them like his children or he like places them as a part of this kind of evolutionary schema.


And the way this is done is like very racial and very awful. But I think it does reveal something about this sort of positioning of the AI genius is supposedly omnipotent and trying to continue on their own, like paternal line.


As Anne Cheng, this wonderful literary theorist and philosopher talks about, race is, the toggle by which these sex bots are both inhuman- they are not quite human- they don't deserve to be treated as human by their creator, but at the same time, race is that which signals to us that they're supposed to be human because, the human is always racialized, it's always gendered. So the human can be broken down in its component parts. So when you're building a sex bot, in order for it to appear human, it must have a race, it must have a gender. So then it gets us thinking, what would it mean to have a sex bot that queered race and gender so that it could be a kind of queer inhuman in a certain way that it wasn't subscribing to gender and racialized norms. And we just need more imagination, we need a different kind of imagination to get us to those spaces.


KERRY MCINERNEY:

Absolutely. I think this is a really beautiful place to slowly bring our episode to a close, which is very sad 'cause I really could complain about science fiction films all day. But Eleanor, why do sci-fi films matter when we're thinking about AI and AI ethics, and what would you like to see change going forward?


What's your key takeaway from this episode for anyone who skipped the rest of it?


Science fiction films are massively popular. They are the big blockbusters, so whether or not you like them, you have to care about them. I did my PhD on science fiction not because I was a massive fan, although I really like science fiction when it's good, and I love the science fiction community cuz they're just like a really nice bunch of people and you get hugs at the end of conferences, but because it's a genre that is widely accessible. It's really pulpy. Lots of people are influenced by the philosophy behind science fiction. So it melds philosophy and literature in a really interesting way, and you get these beautiful genre blurrings sometimes also between detective fiction and romance and science fiction, which is what happens in Italy.


There's so much amazing science fiction written by queer authors, by women authors, from across the world, and these books are not being made into films. No one is spending the millions that it takes to bring these wonderful worlds to life. So I highly recommend that you read Aliette de Bodard's Xuya Universe series.


If you want to see how ancestor spirits and machine and women and different kinds of power hierarchies come to play in this world where geopolitics has been transformed.


So the domination of Anglo-America that we see today hasn't come about in the future, and that makes it really interesting.


There's some amazing books by the Spanish writers Lola Robles and Elia Barcelo, where these female -looking aliens actually make human men pregnant, and they're these wonderful comedies that could be made into fantastic science fiction comedy films.


So I would really urge producers, budding, screenwriter actors to try to bring these amazing ways of thinking about the future to life.


KERRY MCINERNEY:

Absolutely. And for our listeners, we would love to hear your thoughts as well. What's your favourite science fiction film? Which science fiction films do you hate? And what kinds of stories would you like to see told? As always, the full transcript of our episode will be on our website, www.thegoodrobot.co.uk, where you'll find not only the full transcript of this episode, but also a reading list of everything that we've mentioned, including those wonderful stories Eleanor has talked about and anything else we think that you might like to read or listen to.


But until then, thanks for listening to this week's Hot take and we'll see you again soon.






90 views0 comments

Comentarios


bottom of page