 Hello, good evening everybody. It's so nice to be able to breathe and to meet you all here in such large numbers That's a first for this series here for a very long time to have so many people here on site So thank you for coming. Welcome here at Spree Spatia in Berlin Welcome to our viewers streaming this event live on a device near you On Alex TV on the respective websites of the partnering institutions of the lecture series making sense of the digital Society that has been running for five years actually It is a joint venture between the federal agency for civic education Bundeszentrale für politische Bildung in German and of course the Higde-Humbert Institute for Internet and Society but tonight is a very special night in more than one sense because the following lecture the ensuing Conversation with the help of you here on site the audience and at home through a participatory tool called Slido is Called it's the kickoff of a three-day conference of call Of course as you can see called artificial intelligence and the human cross-cultural perspectives on science and Fiction so on behalf of our series making sense I as the mere moderator of these events would like to really warmly thank the Japanese German Center Berlin for co-hosting this very Promising conference on cultural specificities of AI and how they vary in different parts of the world and why it has taken so long To a knowledge these differences. That's the point We're going to touch on pretty soon in the talk We're going to hear tonight above all let me thank Julia Munch the secretary general of the Japanese German Center Berlin Beautifully located if you don't know it in Berlin Zylendorf by the way not far from the main buildings of the FU the Freie University here at Berlin. So thank you Many thanks, of course also to the main curating force behind this program Thomas Christian Pechle From the Humboldt Institute for Internet and Society here in Berlin Central for the beginning of this project and furthermore is one of the visitors one of the quite a few visitors actually From Japan, of course, thank you Watanabe Katsumi from Wasera University in Tokyo Please a big hand for all those people involved making this event possible. Thank you So I'm very excited to start this conference by introducing shortly to you tonight's Speaker and her topic so central to the scope of the calming days of those of that conference How the world sees intelligent machines the world stress the world very eager to learn here more Myself since it is a topping opening up massively in times while its cracks had largely gone unnoticed for so many years Decades maybe even centuries in the West at least. So let me just give you As an anecdote maybe just one more example that ties in with our event of tonight here So let me ask if there are any dancers in the room here Are there any people that? Could not go to clubs or to raves or to anything like that in the last two years. I'm sure there are yeah I see a couple of hands rising here. So What would you say this music actually originated quote unquote? What would you say not in Germany? Although Germany did play a role, you know bands like coughed work No, I can they were called the so-called cow talk, which is very repetitive and the electronic originated in Western Germany, but central no doubt to the Originating music are of course African-American inner cities Detroit Chicago New York even there's many books about the histories of house music techno music and so forth But it took decades until the West discovered a record from India dating back as far as 1982 check it out. You can see it on YouTube actually the record is called ten ragas to a disco beat By a musician. Nobody knew in the West. His name is Jaranjit Singh This record sounded an awful lot like acid house from Chicago Even though the melodic material in some of the micro rhythms were specific to ragas to a very ancient Indian tradition, of course But one answer to this stunning likeness of this music Nobody knew about in the 80s in the 90s even in the 90s One answer to that is they all used the same machines in Chicago in Detroit in New York and in Mumbai machines made in Japan by the way based in drum computers called three or three in 808 and a synthesizer called the Jupiter 8 all made by a Japanese company called Roland I'm quite sure or hope that today would not take decades for the West to see these diverse routes That grew in different parts of the world sometimes simultaneously But of course these machines could hardly be called what we call artificial intelligence nowadays some of them work with Transistors that were deliberately broken actually to get a very specific distorted sound So that's a little of a different thing But maybe it's a small ramp to lead you up to the research of Contadi how and their keynote How to how the world sees intelligent machines she's a senior research fellow at the lever Hume Center for the future of Intelligence at the University of Cambridge and leads to research projects there one is called global AI narratives the other Decolonizing AI in which he explores intercultural public understanding of artificial Intelligence as constructed by fictional and non fictional narratives So we're gonna hear a lot about books and films also This evening since our speaker has also a background in the humanities born in the Netherlands She started her academic education with degrees in film and literature before she got her PhD at Oxford in science Communication her thesis explored the communication of conflicting interpretations of quantum physics to adults and children She's also the co-editor of the books AI narratives a history of imaginative thinking about intelligent machines at Oxford University pressed two years ago and the new one Imagining AI how the world sees intelligent machines the title of tonight's talk due to come out later this year She has also co-authored a series of papers on AI narratives with Steven cave including the whiteness of AI a concept you will Encounter in her talk in a matter of seconds now and after the lecture We are going to start the conversation here on stage Just the two of us for maybe 10 15 minutes before taking your questions from the floor at home With slide out the participatory tool if you're watching the stream, but now the floor is entirely hers Please welcome Kanta Dihar. Thank you very much Toby. Thank you so much everyone for having me Thank you for granting me the honor of opening the conference with this lecture tonight and and my particular Thanks go to Thomas Bachler for Inviting me and to Lena Hinkins for making sure that I got here on time with all my things and who arranged the whole trip I mean for many of us. This has been quite a while that we've been able to do all this And so I would like to start off tonight with a question for you to consider How do you describe AI to a friend? Think about how you would explain what exactly AI is What would your explanation look like would it be something like? Artificial intelligence seeks to make computers do the sorts of things that mines can do Or is it more likely to be in the not too distant future all daily tasks will be carried out by machines Leaving us to become even more lazy and idle than we already are and the Terminator movies will become real Or will it be something like scary robots? So the first of these is a definition by an expert Professor Margaret Rodin who essentially invented the field of cognitive science The second and third and all the others on this slide right now were definitions given by members of the British public when my colleague Stephen Cave and I Conducted a national survey of attitudes towards AI and These kinds of results indicate several of the reasons why I did the research that I'll be talking about today first of all a gap between expert and general understanding of AI second the rule of stories particularly science fiction in shaping those perceptions and third and most importantly These views are taken from members of the UK public and as I'll be showing you today These views perhaps for the best are not universally shared across the globe Before I go any further I should point out that all of the research that led to the information that I'm giving you today was Conducted together with my co-author Stephen Cave who is currently in Cambridge, but who does deserve equal credit So in tonight's lecture, I will introduce views and visions of AI from around the world First I'll be starting close to Berlin and its history With visions of AI in communist states particularly in the 20th century Then of course in keeping with the conference's Japanese German theme I will be addressing real and apparent differences between eastern and western portrayals of AI and Finally, I'll be giving examples from around the world of narratives of AI that explicitly aim to reject colonialist views of the technology But first for some context, who am I to be doing this and why am I doing it? so I'm a senior research fellow as Toby Muller just Mentioned at the lever whom center for the future of intelligence or CFI for short Which is based at the University of Cambridge, but it's a collaboration between four universities Cambridge Oxford Imperial College London and Berkeley in California Now I have a literature background specializing in science narratives as Toby mentioned that includes science communication as you might find in popular magazines or on TV as well as science fiction any kind of story fictional or non fictional or somewhere in between about science and That's how I ended up working at an artificial intelligence research center with a literary studies background So CFI is an interdisciplinary research center Founded in 2016 and it's focused on the nature ethics and impact of AI We've currently got 20 research projects that range from responsible innovation to exploring the relationship between machine and Biological intelligence and we've got the world's first masters degree in AI ethics And our researchers come from a huge range of disciplines In including also machine learning law philosophy, etc And I work within a program called AI narratives and justice which looks at questions of artificial intelligence and social justice So the two research projects that inform my talk today global AI narratives and decolonizing AI originally developed from a project started in 2017 called AI narratives and in that original project We investigated the portrayals and perceptions of AI in the English speaking world So that resulted in that 2020 book AI narratives Which showed behind me and since we also had interest from people developing the technology from industry Policies such as the United Nations, so people who aren't academics and who don't want to go through a 300 page book We also wrote a non-academic report and the sequel to the book Will be out at the end of this year under the title imagining AI how the world sees intelligent machines Hopefully end of this year. It's a contributions to that book contributions from all over the world Which inform the examples that I'll be giving in this lecture So in the initial AI narratives project, we examined the dominant narratives about artificial intelligence In the English speaking West and how they stereotype the technology but also those who build the technology and Those who are affected by it and that tradition was I was starting place because it is Disproportionately influential How influential exactly? Well, one key point to bear in mind is that artificial intelligence was a cultural phenomenon Long before it was a technological one in some cultures visions of intelligent machines go back centuries even millennia like in ancient Greece and ancient China and Such visions keep popping up throughout history ideas of building intelligent machines out of metal out of wood creating life from non-living material and Those visions spread with industrialization Especially until by the 20th century a future with those kinds of machines with being really richly imagined around the world so When the term artificial intelligence was coined in the US in 1956 It was not to name a new invention, but it was to express a determination to realize an ancient fantasy and Some would say that AI is still a cultural phenomenon and not a technological one Some would say that for all the innovations in computing all the hype in industry and policy No existing systems deserve to be called truly intelligent But others argue that we are surrounded by AI It pervades our daily lives through our smartphones the online services we use and the hidden systems that govern us now It's hard to think of another technology in history about which such a debate could be had a debate About whether it's everywhere or nowhere at all That it can be held about AI is a testament to its mythic quality The amount of storytelling that is part of this technology Now of course innovation in real digital technology is actually happening It's really rapid and it's constantly interacting with that ancient set of stories of intelligent machines So this cultural background shapes what motivates funders and engineers how products are designed Weather and by whom technologies are taken up how they are regulated and so on Academia industry media policy They're all interwoven and they're mutually influential So let's give an example From the Western Canon to show really how the dominant stories and the dominant technologies Connected and that is the 2014 film Transcendence a Hollywood film which tells the story of an AI expert will cast her played by Johnny Depp Who has his mind uploaded into an AI system? Now during this film, so this film was not a success. It completely bombed. Okay, which is why you probably haven't heard of it, but During the film's opening weekend three very well-known scientists who you may have heard of unlike this film Stephen Hawking Max Tegmark and Stuart Russell published a Huffington Post article titled Transcending complacency on super intelligent machines and In that article they argue that and I quote from the article as the Hollywood blockbuster transcendence Debutes this weekend with Johnny Depp Morgan Freeman and clashing visions for the future of humanity It's tempting to dismiss the notion of highly intelligent machines as mere science fiction But this would be a mistake and potentially our worst mistake ever So in this article these three really influential scientists use this film to convince policy makers and other groups of the importance of AI and That same month one of those authors Max Tegmark Founded the future of life Research Institute which aims to ensure that tomorrow's most powerful technologies are beneficial for humanity That Institute is partly funded by Elon Musk Who is a real-life technology magnate and pioneer of AI driven cars Musk also appears in a cameo in transcendence as an audience member of the lecture that Will Castor the fictional AI engineer gives about AI in addition Morgan Freeman who plays an AI engineer in this film sits on the board of the real future of life research and Institute alongside Elon Musk So the film therefore perfectly reflects what I call the Californian feedback loop this entanglement of Hollywood of academic research of industrial production of narrative and the fight to shape the future So stories like transcendence literally co-construct what AI is Understood to be its embeds or disputes existing attitudes and approaches It's creates it directs funding streams and all of that comes out of one very small part of the United States But crucially those attitudes those approaches are not the same around the world They are shaped by the particular histories philosophies Ideologies religions narrative traditions and economic structures of different countries and cultures and peoples Transcendence is a product of the US and it's full of well-worn Hollywood tropes Which are to some extent more or less problematic there is AI as the ultimate technology and the ultimate solution to all problems the mind uploaded Johnny Depp solves pollution the energy crisis disease But at the same time AI as the ultimate threat to humanity because mind uploaded Johnny Depp does go rogue the reduction of an individual to data and computation mind uploading and so the possibility of digital immortality and The lone male genius scientist and a subordinate female Who must in some way sacrifice herself as she literally does in order to make sure that? The AI is is to stop in time before it takes over the world Now individually all of these ideas are not unique to Hollywood each can be found elsewhere but collectively they form a really distinctive mythology of AI in America and There are some serious problems with the fact that this group of narratives keeps Reinforcing each other because it's quite a small set of ideas I mean what you see in nearly every story about AI to come up from this area is for example The extent to which portrayals of intelligent machines are anthropomorphized are made to look like humans and Correspondingly they are gendered they are racialized and AI is made that looks like Johnny Depp and the mythological and imaginative tradition that leads to Current conceptions of AI is quite different to the way other technologies are being portrayed other technologies are not Personified in that extent intelligent machines are never portrayed as Just tools they have always been agents Making them look like humans Implies that they can do all these things that humans do but also that they take on human roles in Society that they follow human social patterns And so for example, we imagine AI servants AI soldiers AI cleaners And and not just those Human jobs, but also social roles in relationships friends lover child parent Which means that you have to position the AI in very complex social hierarchies And so while these machines are taking on human roles are shown in fiction that shows them taking on human roles They are exaggerated in some dimensions and flattened in others as Soldiers they're hyper masculine as lovers. They're hyper feminine They're supposed to show initiative and intelligence in their roles, but at the same time people hope that these robots are Unthinkingly loyal slaves and if they are not then it is a problem and they are depicted as a threat to humanity And Now of course the concept of robot slaves is a highly racialized one But here comes another problem of the stories to come out of the Californian feedback loop Intelligent robots are only very very Rarely coded as racially Any kind of person of color And on the contrary they are overwhelmingly depicted as white both in ethnicity and just in Color as you can see here So that is what we examined in the paper the whiteness of AI that Toby mentioned So in that paper we examined the way in which AI in the West is used to create a kind of white Utopia from which people of color are entirely absent and the consequences that has for public perceptions and the implementation of contemporary AI technologies So there is a stereotype of the kind of products that are created the intelligent machines themselves, but these kinds of narratives also perpetuates and exacerbate stereotypes and expectations about the creators of these project products the AI scientist or developer So as you can see here some examples in addition to Johnny Depp earlier The dominant AI narratives in the Anglophone West tend to be literally the product of the white male Imagination in that the great majority of those who have created these dominant imaginaries of AI are educated white men But at the same time in those stories white men create white robots So even in our imagination AI is the product of the white male imagination and We might call the cultural construction of the AI engineer the white male genius trope Which is a strong one in in all of these films from which I'm showing stills behind me and That we think is a very strong factor in contributing to in real life the low number of women able to enter the field of AI and The hostility they face when they are there for not fitting in culturally Now we're currently actually near the end of a project analyzing a very large corpus over a thousand films and 10,000 episodes of TV shows looking at how AI scientists and engineers are gendered and racialized in Anglophone film and television So those are some of the themes from the Hollywood hegemony and some of the reasons why we might want some alternatives so is It inevitable that we imagine intelligent machines this way we didn't think so and We wanted to find out exactly what alternatives there are for four reasons at first AI is now a global phenomenon while the term as I said originated in the US and much of the technology Continues to be developed there These technologies are now being taken up around the world and other countries are scrambling to develop their own AI industries now each will do so informed by their own mythologies of AI their own sets of stories and ideologies that shape their expectations and anxieties around what that technology can be and So understanding how AI will develop and how it might differ in different parts of the world Requires an understanding of those many places those many sites in which its story is unfolding But also the debate around how AI is developed responsibly how it's governed has been dominated by Anglophone actors and It is starting to change more countries are developing their own AI strategies, but they are entering a space that is already shaped by Assumptions from the English speaking West So there is a risk that efforts to regulate real-world AI will fail as they are insensitive to different cultural contexts or as that they will impose solutions that Unknowingly or inadvertently prejudice some traditions So it's extremely important to develop a better understanding of this diversity of views of what AI should or could be And of course we hope that that comparative approach will shed new light for people Whether here or elsewhere. I mean Seeing where other cultures share or differ in their approaches to AI Gives insight into the forces that shape those traditions So for example when you've got a range of narratives from Capitalist countries you could see what do they have in common or what do countries that have histories of Colonialism have in common and how does that shape their stories about AI or on the other side Countries that have been colonized So for example, I've been looking at Anti-colonial or decolonial AI narratives from Latin America, India, Sub-Saharan Africa and Hawaii Which all show that while resistance takes many shapes shared themes resonate across continents So from the platforming of non-western knowledge and forms of knowing with respect to AI to critically reflect on What is considered advanced versus Backward to reappropriating technologies from the West for purposes for for new purposes new art forms and For using these kinds of stories specifically for post-colonial nation-building But we see also narratives about catching up to other countries from South Korea India and Russia and Narratives have historically resisted and supported a really broad range of ideologies from communism in mid 20th century China and the Soviet Union and Italy to neoliberalism in Chile to Technocracy in Singapore and I'll be touching on some of those Further later on but also Each cultural perspective is limited and particular and it privileges some within that culture and prejudices us so while the English-speaking West is inflected by ideologies of racial and gender and class hierarchy People have been calling for new imaginaries of technology Because each culture has these kinds of blind spots and so only by crossing over and comparing notes Each culture can see what their own lacks are of what their own Blind spots are the bits that they are missing from their narratives Now of course The limitations that I've just pointed out are not solved simply by you know adding some Chinese works to your reading list are consulting an indigenous person or Mentioning Ubuntu ethics at a workshop on AI regulation But what's been happening so far is that these kinds of narratives have not been given the light of day at all and that just means that the kind of thinking Around these issues has been a lot shallower than it could have been simply because The same stories that you have already heard are repeated again and again So it was with all that in mind that we started the global AI narratives research project in 2018 it was in collaboration with nine partner institutions on six continents and We aimed to well understand and analyze how different cultures and regions perceive the benefits and Risks of artificial intelligence and we did so through convening a series of 20 workshops across the globe between 2018 and 2021 Which we moved online at the start of the pandemic and through that we built an international network of experts on Portrails and perceptions of AI beyond the English-speaking West With relating that to pressing questions of real AI ethics and governance right now So all that was our motivations for doing this work and now I'll move to sharing some of the results and As I mentioned, I'll start relatively close to home for my Berlin audience So we held workshops in Russia China and the Czech Republic And so we've really frequently encountered these narratives of AI in a communist context now scholar and science fiction author Anton Perushin has pointed out that Soviet narratives about artificial intelligence presented mainly ideas of very trustworthy AI's in the service of humanity such as in Ivan Efremov's andromeda nebula from 1957 which explored a communist utopia in which an AI acts as the logical Overseer and decision-maker although there are still humans overseeing the overseer and Those kinds of narratives in Russia really the ones about the subservient or assistive AI influence people's perceptions of what real AI could be like so Until the 2010s the ways People expressed what they wanted from AI was a Terminator in the service of mankind Which basically means that they demanded of chatbots that they must know everything not have any emotions Perform commands answer questions not the bit where it travels through time to kill people But recently this terminator in the service of mankind few has changed people nowadays Don't want those stereotypical movie AI's anymore, but likeable human-like chatbots that don't know everything and that can make small talk and central to those kinds of ideas is the concept of As Anton Perushin calls it the Russian soul which analyzed the notion of a Russian national and cultural identity Questions of what it means to be human whether an intelligent machine could be considered human by virtue of its exhibiting Traits that evidence having some kind of soul Were really closely related to questions of what it meant to be a Soviet citizen So in one Soviet film, I'm not making this up The most important test for the AI to prove that it was human was whether it could drink vodka But also in a very popular TV series from the 1980s called the adventures of electronic Which was based on a children's book. There's a robot Electronic that looks like a boy and aspires to be human So electronic decides to go to school with other children and learns how to laugh and how to cry And so rather than having a whole range of evil robots Like in the Western tradition this very popular little robot is one that is funny But the most influential AI narrative to come out of Eastern Europe is of course Karel Chappek's RUR Rossum's Universal Robots a play from 1921 and that's a story of artificial workers rising up against humanity and exterminating them and It's a very clear metaphor of the oppressed underclass violently rising up against the bourgeoisie That play was a huge hit across the world not only in Europe where which recently had had such violent revolts all the time But also from the US to Japan and in the USSR It was adapted by the famous writer Alexei Nikolaevich Tolstoy as the revolt of the machines So a version that is even more explicitly aligned with communist ideology And in the mid 20th century cybernetics the first kind of science that Investigated well that was served as a predecessor to what's now called artificial intelligence became very interestingly entangled with Soviet ideology it was first developed in 1948 by an American scientist Norbert Wiener and Initially his work was banned in the USSR His ideas were considered a threat to Marxist Leninist philosophy of science But in the post Stalin era Cybernetics had a complete revival in the USSR It was seen as a way to break away from the hard labor that defined the Stalinist era So to create an opportunity for a really egalitarian and post work world so luxury Luxury space communism And I should mention the role that US science fiction in translation played here For example Isaac Asimov who is one of the great shapers of the American AI canon While he moved from the Soviet Union to the US at age two He really took pains to depict himself as completely American and Back in the Soviet Union He was perceived as American when his works were translated into Russian a couple of decades after they first came out and actually as my colleagues Angelica Solovieva and Nick Hineck have pointed out as Asimov's work Resonated really strongly with those Soviet Imaginations of intelligent machines that try to be human and try to be nice and funny So they write Hollywood often featured terrifying killer robots such as in the Terminator Whereas Soviet people usually imagined robots as almost Indistinguishable from humans in both appearance and behavior as we saw when we looked at the boy electronic So paradoxically they say it was Soviet rather than American robots Which were designed to naturally observe Isaac Asimov's three laws of robotics The laws that he became so famous with in all his robot stories from the 1930s to the 1990s So next I should look at China as a bridge between my first topic AI in a communist context And my next the differences between East and West Now China has several ancient philosophies Confucianism Moism and Taoism that have for thousands of years expressed views on technological advance and Whether developing intelligent machines is a good thing whether developing specific kinds of machines is a good thing Nonetheless as the scholars Chiang Bai Chun and Qian Miao have argued Real technological development often flew in the face of sage advice. So for example while Taoists warned against the development of advanced technology for warfare in purposes weaponry developed really rapidly from the spring and autumn period, which is around the year 770 BC all the way through to the Han Dynasty which ended in the year 220 so in spite of Philosophers writing you should not develop technology for this and having really advanced essentially technology ethics And in fact the technology developed at that time far beyond what anyone else in the world was capable of Moving forward to the modern era So in China following the Chinese Civil War from 1945 to 1949 Science fiction literature served really a political role Our colleague Wu Yan explains that in the science fiction of the early people's Republic. So this was under chairman Mao Writers were consistently concerned about the issue of AI technology replacing human labor Humans being put in replaced by machines Nothing new and nothing solely reserved for Western imaginaries So opposition to the replacement of all kinds of labor by intelligent machines was the dominant idea in science fiction during this period Most stories argued that AI would create lazy people Reduce human motivation to learn and set society back But after Mao's death Chinese science fiction actually became more heavily influenced by the Soviet stories that I've just been looking at It wasn't a straightforward influence because Soviet works were during the Mao era Often criticized or banned for not properly aligning with Marxist values as they were interpreted under Mao and Like in the Soviet examples Science fiction in China now began to feature stories about artificially intelligent machines built to solve problems and Being helpful and obedient servants to their creators So come going from the more fearful and worried AI going to take our jobs stories to the more positive and and hopeful stories about living with AI Continuing the theme of looking at eastern AI narratives Let's look at Singapore where governance and technology in the late 20th and the 21st century are Really closely intertwined and Singapore is really unique in that technology is an integral part of the dominant and official utopian visions for the island and There's a long tradition of imagining the island state of Singapore as a potential Utopia that's a tradition that began with the British and continued by Chinese invaders and settlers But sing lit as Singaporean literature is known tends to be dystopian with a very specific function To provide an alternative to the narratives promoted by the government So those who don't agree with the kind of utopia that the government wants write these kind of narratives now to Japan Japan is of course and I must of course Say that I am not the expert on Japan in the room here today or this week But Japan is perhaps most famous for being different from the West in imagining a future with AI There are many ways in which this is absolutely true There seems to me less concerned about AI taking our jobs than there is in most Western countries As my Japanese colleagues have pointed out partly due to a very rapidly aging society in which there are simply Aren't enough people of working age in Japan to do all the work that needs to be done But this is also one explanation why Japan is less concerned with themes of robot rebellion and Just like in the examples from China and Russia There is a tendency to portray intelligent machines as friends as helpers as extensions of humans instead of you know murderous and rebellious others But the most widely known difference between Japanese and Western ways of imagining AI is based on a difference in philosophical traditions of which the Japanese tradition of animism emphasizes connectedness between different kinds of entities such as human animal and machine But now what I find a really interesting to learn is my colleagues Daniel white and hero for me Katsuno have argued that the role of animism in this perception of AI is a lot more political that he might at first think They claim that animism was reinvented as a cultural model to define the relationship between people and robots in Japan during the Japanese robot boom of the late 20th century For a specific reason While the theme of human-robot partnership did come to dominate the technological imaginary in Japan It often did so in conjunction with practices of representing technology in Japan's modernity as a symbol of cultural distinctiveness in opposition to the West So in other words Japanese philosophy was being deliberately foregrounded as a way to emphasize how unique and different from the West Japan and its technology was It was animism as a unique selling point to put it much more bluntly that my colleagues have done in South Korea Has many of these aspects in common with Japan, but I also found there are some notable differences For example, there is some concern about automation related job loss Even though the population is aging almost as rapidly as it is in Japan But the most important difference is due to what has been called alpha go shock So as many of you may have heard in 2016 the British AI company DeepMind Set its go playing program alpha go against the South Korean grandmaster and world champion Lee Sedol and The computer one that was a milestone for AI development because it showed the success of a new way of training AI But it was a huge shock in South Korea for two reasons first, of course the shock to tradition and culture that comes with the South Korean world champion It's such a nationally important game being beaten by a computer But second the realization that this computer was not made in South Korea or even anywhere in Asia But in the UK so that has led to a catch-up narrative the a story that Other countries are ahead and South Korea must catch up in the development of AI technology and to that purpose since 2016 this narrative has led to vast sums of money being invested in developing the AI ecosystem in South Korea So the third and final theme I'll be touching upon tonight is the idea of decolonizing AI Particularly through the lens of narratives of AI that explicitly aim to reject Colonialist views of the technology which is a theme that frequently emerges on the one hand in narratives about AI in the global south and On the other hand among marginalized populations in the global north I'll show three kinds of responses to neocolonial dominance in the field of AI absence resistance and reimagining First absence our work in the Middle East and North Africa about which we co-authored the report pictured here Showed that well many local Thinkers and scholars Claimed that Egypt in particular was a so-called AI desert that there was no Development of AI technology ongoing in Egypt very little in North Africa Nor were there any notable films or literature or non-fiction works stemming from the region that portray a future with intelligent machines in any way that could compete with The idea that everything is being imposed either from the West or from Japan both technology and stories But that's not the only way of looking at this region They call it the AI desert, but a desert is not empty and lifeless Particularly on the Arabian Peninsula Nations are developing their own hybrid of Western technologies and stories with local approaches So one example is the Ibn Sina robot pictured here, which speaks Arabic But not everyone is equally happy with that hybrid. Some call this move Self-orientalism using Western technologies with aspects that the West would consider typically Middle Eastern such as a robot wearing a Thobe and a kufia or science fiction stories featuring gins So that is absence, but so AI narratives have also been used as an explicit form of Resistance and my next example focuses on AI as a form of anti-racist resistance in Brazil The scholar and Edward King has pointed out that social media in Brazil is a double-edged sword on the one hand as We're familiar with around the world Its algorithms drive up and encourage the clustering of like-minded people including Racists and spreaders of hate speech on the other hand. It has become a platform for resistance and in Brazil people Resist and provide counter narratives using the aesthetics of Afro futurism Which is a movement that focuses on the experiences and concerns of people in the African diaspora Through technology culture and science fiction and a wide range of black Brazilian artists Use technologies to create new ways of imagining a future in which they are in charge of rather than victims of technology So here's one example on the slide Still from a 2020 video by the artist fit or in a crib And finally I will address ways of re-imagining AI of Thinking about what AI might look like in the future if we don't follow the dominant narratives and predictions So three years ago a group of indigenous scholars artists and thinkers from indigenous communities all over the world Started a conversation on how to center and engage their voices in the future development of AI and one of their approaches in the words of the scholar Jason Edward Lewis was to imagine AI as a Helper finding the middle ground between the Hollywood stories of He said as he said Blade Runner, which is AI as a slave and Terminator, which is AI as a tyrant and instead Imagining a middle ground in which AI and humans are in reciprocal relationships of care and support he and His co-authors wrote an amazing position paper in which they offer a series of vignettes that show what that might look like Which is from a multi-sensory Computing device based on the traditional bags of the Algonquin people in what is now eastern Canada To a poem about a child raised by three AIs that are developed from Hawaiian and black foot values Which is illustrated here on the slide So our one most Important takeaway from five years of researching AI and fiction is that the extreme and homogeneous Narratives that are dominating the Western discourse are not the only way to imagine a future with intelligent machines They have been dreamt of across the world in very different ways indeed Thank you for listening tonight. Thank you so much Conta for this very global outlook indeed and the histories of how we perceive AI and so many different regions and how to counter that whiteness of AI you were talking about one of the Probably strongest narratives in film and we talk about AI in the West At least is by a British film director Stanley Kubrick 1968 2001 a space Odyssey it was called For those of you who haven't seen that film it Was dominated actually and the voyage to Jupiter on a spaceship by an AI which was truly an AI which was you know Like a singular singularity almost called how which of course was a word play with IBM then the dominant Company when it came to computers how just preceding you know age preceding I a preceding B L preceding M in the alphabet so HAL IBM that was the thing there and how was Superhuman, but if you you know if you see that dialectically it was very human so to speak it was How you called the anthropomorphic almost it didn't have a body really but how it felt how it spoke It was very much structured Like a language so to speak so would you say that anthropomorphization? Which is something that is widely discussed in AI is actually a western concept or is something? Is it something that you see at various places like in the Soviet literature that you showed us? It was anthropomorphized very much even more so than in the West although it had different traits But is this like something you would call to be universal when we talk about AI and I'd say anthropomorphism is quite universal Yeah We have seen it all over the world and even when you look back at these really ancient ideas of imagining Intelligent machines it is actually it starts with anthropomorphism That's really how people imagine going to a whole new level of creation of technological development so say you have a Someone who is really good at building new technologies. How do you? Write a story in which you prove that this person is absolutely extraordinary How do you prove that this is the best? Craftsman technologist developer of a generation almost godlike well you make it usually he imitate life and So there are stories about mechanical animals so artificial animals, but of course and Humans all like to think of themselves better than any other animal. So Again, the next level is making an artificial human And so stories of artificial humans of humans made of gold and bronze and wood and metal are Some of the oldest ones that we have about AI so it's it's definitely I guess Biomorphization so making it look like any kind of life form and within that anthropomorphization that is the most important idea behind The old stories that now influence real AI even though we know that it's not going to look like a human anytime soon We had a guest and then in the series making sense of the digital society I think at the beginning of the pen of the pandemic Joanna Bryson and as she talked length about saying that Anthropomorphism is actually one of the biggest traps when we think about AI and it's Very hard to communicate this with the broader public that this is not what current AI discourse actually is about not on the Technological side what you refer to at the beginning of your talk. There's no true AI It's machines that are being built actually and not singularities, of course So how to counter that in a pop cultural Sort of way because there's many traps some of them you hinted at one of them is of course racism that I mean we anthropomorphize AI in the West that we end up with a lot of weightness on robots even and Just how we picture AI how to counter this Yeah, so I completely agree with with Joanna on this that imagining artificial intelligence as Looking like a human acting like a human just leads to all kinds of problems Partially because of what the technology is capable or incapable of right now and and partly about what kind of rules we want for that technology and so what I do find very interesting is the kinds of cultures where we can learn from that have healthier Relationships to the world around them compared to cultures that have very hierarchical ideas about Humans are at the top and the rest of life comes comes below us on the ladder or even My type of human comes at the top and then there's the rest of the humans and then there's the rest of the animals And and so when you think about for example these in indigenous Narratives which provide a much Healthier way of imagining that it's perfectly possible to live and work with And a kind of intelligence that's completely strange a kind of intelligence that is unlike Anything that you've seen before and that's certainly not like a human That is a much better way to start thinking about it because if we ever make something intelligent Out of silicon it is not going to look like or act like or think like a human And some traditions are just much better at imagining that very interesting now in your research For many years now have you encountered any kind of Change or something that sort of spurred change in cultural views on a let's take the example of Japan of course that you refer to it that you said there is less concern and Japanese history with the rebellion of machines that they're more perceived as helping hands to master daily routines to do care work for the old for the elderly or for the sick and so forth Could that change what would have to happen that those narratives sort of evolve develop change Turn around have you seen examples like this happening elsewhere? I think that a big change has happened after AI technologies Started to be called AI technologies again from sort of between 2012 and 2016 so which is Before that we had what is known as an AI winter in which very few people Talked about AI very few people said they were working on it It was an extremely uncool topic and there was no money in it at all until in 2012 suddenly Image recognition Improved massively almost overnight by the introduction of new machine learning technologies And so that created a new AI hype wave that was still riding to this moment and especially as several major improvements and transitions in AI happened, but the problem was that So until a few years ago if you wanted to talk about this new technology that people called AI What you had to fall back on was the most recent big AI Narrative that you could remember seeing in the cinema, which was probably in the West the matrix or the terminator and so I Don't know if it was the same here in Germany, but in the UK They're literally slapped terminator pictures on everything and the government put out a report on the UK's economic future with AI bam terminator picture Article on FinTech Terminator picture an article on an invention of drones that could fly backwards terminator There's a certain logic he became governor of California, right? Yes, well he was known as the governor and So the problem was that you had these old stories and new technologies and they didn't match up But what we did see over the last few years particularly and sometimes Authors have been explicitly commissioned to write new stories to create new narratives Other times there has just been so much criticism of these old stories that people have been inspired to create new ones And so over the last three years or so there's been an explosion of AI narratives that are a lot more Exciting in terms of how they respond to contemporary technology and sometimes you do see that people then start borrowing from other traditions I think that the children's Animated film big hero 6 is an interesting one because although it's an American animated film it is Very strongly influenced by Japanese film Traditions, it's actually that the whole aesthetic of it is somewhere Literally halfway San Francisco and Tokyo in a fictional city called San Francisco in which a child builds a Robot a care robot and the whole story is about how this robots How he and this robot work together To get revenge on the baddies But also how this robot becomes a true supportive companion and how in the future he can see These care robots really curing people and and so it has a very positive message And that I think is a direct influence more from the Japanese side. Thank you for that example In a paper you cited with your co-author Steven cave You discussed three possible Solutions to sort of well, let's say plainly to fix racist AI right to make it less white to make it less dominantly white You said three things avoiding anthropomorphization That's what we just discussed now to more things explicitly critiquing racial role typing and representing powerful AI as non-white so I would think that the latter two critiquing or critique of Racial stereotyping of course in representing powerful AI as non-white should be Easy, I mean that should be possible. That should be possible to fix right? Do you see as much progress as you're hoping for on that front at the moment and where there is some? progress there are some Especially in TV I find that in TV progress comes much easier than in film perhaps because TV can introduce a much bigger cost so that They can have multiple people of color in these major roles But also have white people so that they don't get too many complaints Whereas in film if you put in one person of color as the lead then they get a lot of complaints Which is now the case with Percy Jackson films and before that with Harry Potter and whatnot So yes, I have seen progress happening. So some examples are Westworld where the two main AI characters The three main AI characters. Oh, do I do I spoil this? It's been 2016. It's been a while So there is one white woman one black woman and one black man as the main AI's but also the series humans which is a British TV series about AI which has the British Asian actress at Gemma Chan in the lead And so especially in TV. That's yeah, I'm very excited to see that happen When you say TV that includes of course streaming services, right? Like a global companies like Netflix and Amazon Yes, HBO and so forth. Yes, and actually Amazon No, Netflix has made now a film called outside the wire which has Only one AI in it and that is an AI played by a black man Anthony Mackie who you might know from the Avengers So and and I think that is the first one I've seen coming out of The US with a single black AI in the lead Are they doing it because they know so much more about data? The streaming services to of course, that's their business model building recommender systems that work Yeah, so so so you mean they develop the film in order to respond to the demographic that wants it I probably I think so. Yes Yeah Which which is very cynical take on it, but at the same time it also shows How big this demographic is and how big the ask for these kinds of films is that you can throw lots of money at it and it shows how How underserved that massive demographic is and that demographic is of course not just you know black men who Look exactly like the black protagonist, but anyone who might enjoy this by the way Really interesting film also because it happens. So it's it's a film from 2020 or 21 and It's set in Ukraine where there is a very strong border tension with Russia and and it's set in well the near future because Of course the main character is an AI and so the Americans come to intervene in Ukraine So it has become newly extra relevant One last question before we open this up here, of course to you and at home through the slido Participatory tool the dreams of or imaginings as you call them often of intelligence machines are very old As you showed us not just decades centuries of thousands of years in some cases If you define AI the way you did it at the beginning of your talk So how to how to build new machines that sort of Clash and certain values or instances with very old beliefs and narratives when I say okay respect local tradition or regional tradition or whatever, but Do it with less gender stereotyping for example what to do there? Yeah, that's a really interesting one and I think and it's it's just Really important for developers of technology to bear in mind these kinds of histories and and expectations that you will face when you try to deploy technology anywhere is The idea that people have these automatic assumptions about what it will do and that need to be addressed indeed explicitly That need to be addressed And that could be in for example through what do you call your technology? For example, some people are walking back from the idea of calling everything artificial intelligence because it might be misleading because people Rely too much on it thinking that it's a lot cleverer than it really is and In in other cases People decide to make their robots look more or less like humans depending on How much people want their robots to look like humans and again what kind of expectations they have of of robots and Human like creatures in their house So so yes, and I think in some cases this just has been not given enough attention to and that's where you see these kinds of Clashes like you got with the British responses to what do you think AI is? Where there's this massive gap between what it really is and what people think of it or want from it or are terrified of Speaking of which I think this chair here is sort of Made by a German AI that sort of respected lesser height of Swiss people I'm kind of drowning in this thing here So it's kind of hard to look around but I think there's a microphone out there somewhere and we're taking questions from the floor Now, please raise your hand. We only have one microphone and one guy doing sports Yeah, there's one right at the column Okay, well, thank you for your presentation. I have one question regarding your rules on how to make them less less ratio The you mentioned that it would be better to make them not look human and then you mentioned examples of movies how they were played by Humans that were they were black or at least non-white. Is there a way you could imagine of Reaching both goals at the same time or is it a necessary step to reach? That through non-human way That's a really good question because it's a dilemma that people who work on Human AI and human robotics interaction really struggle with because on the one hand if you if you have a Robot making it look less human avoids all these problems of gender and race But at the same time making something look more human has all kinds of positive effects as well so For example trusting something more than some really clunky gray shiny boxy thing. I mean Say the machine is meant to help Lift patients who can't walk out of bed and put them into their wheelchair You know if it looks slightly More like a human or more like something friendly with arms instead of I don't know something that just looks like Like it comes straight out of a car factory that can really help put patients at their ease You don't want it to look too much like a human. So this is a really big problem and an additional problem is one That's called mirroring which is that actually people particularly from Minority backgrounds or from oppressed backgrounds are more at their ease when they feel like they're communicating with someone like themselves So for example, so there's been experiments where For example, African-American people were put in a chat room with a chatbot that identified as A white man or a white woman or a black man or a black woman with it, you know, just a completely Like artsy looking drawn avatar But it helped They felt better understood when it was presenting as someone like them and so here you get this this this problem of Yes Making things look less human is a great goal to solve some of the problems but sometimes you need to make things look or sound human and Then having a wider range of representations is of course much better than having this very narrow set same also applies to Voice-operated assistants. I mean they there's so much research that has gone into what Siri and Alexa and and the like sound like in order to make To make people feel at ease interacting with them just and and not feel concerned that That it sounds too cold or aggressive or anything I'm still amazed at how Uniform those systems sound. I think you have like with the Siri with Apple You have the choice of a female and the male voice. That's it, right? Where's there more so so originally Siri and Alexa and Microsoft's Cortana all launched only with a female voice female voice only that's right Yeah, and then there was a lot of backlash and saying Look, that's all very nice that research that shows that people are happier bossing female voices around But actually isn't the reason for that a long history of people being really comfortable bossing women around Rather than men so they added male voices and now they've actually got a huge range. They've got so for example in English They've got you know, British Australian Standard American, but also African American vernacular English And yeah, they've really they've now got Indian English. They really improve their range We have that in Germany. Do we have a Bavarian Siri? I Don't know. I'm not joking might be but but do you have German Austrian and Swiss Siri for example? No, I don't think so Oh, it's this learn would be really complicated. You'd have to have a new Siri for every five kilometers So you don't want to do this. This would be just too messy Yeah, but let's take another question from the floor, please speak up. There's one in the back there Thank you very much for a nice talk. So I'm curious about also about Difference of the generations and we are also researching in the Japanese people and People acceptance of the AI and robot is very different from the generation. So I hope I'm curious about yeah it's kind of generation changes also Same as a different country So it was a question about the differences between different generations Yes, yes, it's a great question and I know that in in many countries that differences between For example the generations that grew up with stories like the Terminator and the Matrix seeing them come out and going to the cinema and Seeing these be a really big part of daily life have shaped expectations and There is now actually a really interesting development where people are Growing up and starting to work as AI developers as computer scientists who just Don't get the Terminator because you know the terminators from 1984 second one 1991 bunch of them after that most recent one 2019 but nobody watched that Flapped the last four ones. So it is very much the Terminator is starting to be seen as a last-generation thing and I'm quite excited because as I said, there's a whole lot of new narratives coming out That presents AI in a different light But also This new generation will be able to be exposed again to so many more Narratives from around the world that are now being platformed and so so so for example only 20 30 years ago, it would have been a lot harder to For Western audiences to consume a lot of Japanese Media like manga and anime Whereas now it's just out on Netflix. So I think that the crossovers are really I Think improving Contemporary attitudes to AI Although at the same time people are growing up with a sense of well actually is AI going to somehow threaten my future Although I'm encouraged by the fact that that doesn't seem to be a really strong narrative in in many parts of the world And not a lot among the youngest generation It's usually people who have been in in in employment for a while who might find it difficult to have their To have to change their career midway through who are most concerned about AI taking over jobs And at the same time there are some narratives Periods in the books that have not survived so to speak that might be interesting like take a Blade Runner Do electric Do Android stream of electric sheep was the book called by Philippe K. Dick before it came a movie There's this device in there that is actually nothing like an Android I forgot what it's called where people tap into this sort of syzipus Mythology and do something which is not very anthropomorphic so to speak was left out in the film But is still there but let's take another question before we look at Slido Now I don't see everybody because I see that this is the chair again and the column I don't have Panavision here, but let's take one from this side for a moment and then switch over, right? Let's just I'm sorry And then we'll switch over and they will go to Slido. Yeah, thank you for your patience Thank you so much. Thanks for that fascinating talk. I have a very short Question which is why no South Asian science fiction or maybe you mentioned it a little because there's very interesting stuff Precisely in relation to the wonderful point you made about the relationship between Cultural attachment to things we are here in the Japanese German Center So Japanese tradition of Kintsukuro Iwabisa be paying attention to things not throwing away things and also in the Indic philosophical tradition the relationship between things with consciousness Chinmaya Chit plus Maya and Mrin Maya things which are objects without consciousness So you talked a lot about philosophy of science in the Chinese context now I was wondering if you'd sort of see a connection a continuity between this in the mind-body object continuum. Thank you Thank you so much. And yes, I must say there have been Many traditions that we did look at but that I was unable to include tonight I mean we looked at literally every continent mainly because when we had a workshop in Chile we had representatives from indigenous groups who Technically live under the Antarctic Circle. So we had all eight seven continents taken into consideration. So on Indian philosophy and and and cultural thinking about AI actually in our forthcoming book we have a really interesting chapter by Upamandu Mukherjee who writes about Mostly the post-colonial attitude of India to AI and how while it is strongly influenced by by by British Colonialism and storytelling much more so than by Indian philosophy what they do with it is really interesting is is thinking about AI in in these really original ways in the mid 20th century, which which means that these stories were About robots about relations between Scientific development and how we should treat robots were really present in in in Various languages in in science fiction magazines in India at the time So which developed Not not not to say independently, but more as a pushback against the sense of English superiority about both their own science fiction and their own science and Basically with the attitude but look we can do this too. We do it completely differently and Thank you for asking that So we'll switch over before we go to Slido If you could make it short, we would benefit a little bit because we're running late here. Thank you Hi, I would be really interested if you have seen any attempts our trends where AI is not Depicted in an embodied manner in science fiction because I'm always thrilled from this link of robots or humans and AI Whether it's not existing in reality that much Yes, it's it's really hard, but there are some stories where at least Attempts made to not make it look like like a robot with two arms and two legs And and and all of those come from after the computer was invented It was a thing and so you have you have suddenly this whole rush of Stories about evil computers like how 9,000 or about computers that can control the world in positive ways such as the Soviet one that I mentioned and But it's it's really hard to Write an interesting story About something that is not in any way human like so for example Hell had to still have a voice It still had to sound like a human because he had to be a character in a story And so it's more it is a problem of weighing What a story requires with How abstract you can make it which is again another limitation that means that Stories about AI that we already have are not always useful ways of describing what AI technology is like Okay, one last question, please before we Move over. I don't I can't see you. I'm sorry. I can Nice presentation. Thank you. My question is about AI ethics Do you believe that each country should have its own ethics AI ethics based on the way of understanding of AI or There are like a global AI ethics that should be applied globally. If so, how can we apply this? Thank you. That's a Really difficult question to answer and one that I think has been asked Over and over especially during the 20th century in different contexts people have different philosophical traditions different understandings of what ethics means and How do you build? global systems and networks that are while they are Take into account somehow those differences This is a problem for example in the context of human rights Where well, there has been a really strong attempt at making those universal But actually and and human rights lawyers and scholars have started to point out where in some traditions They just aren't sufficient And so that is also something that the field of AI ethics is struggling with because you can't have Different AI ethics in different countries. It's such a globalized technology. It's made in one place and applied in another I mean a place like Facebook I don't know how many have you know of you have now given up your Facebook But it did at some point have literally half the world on a platform so It's an ongoing question Let's look at Slido Where is The person I there you are. I'm sorry. I was looking in the wrong direction. No, don't worry I have one question. How is religion affecting AI narratives? You talked about the absence of AI narratives in some regions Could religion be a reason for it? The interaction of religion with AI is is really interesting so for example Did you know that Vatican City is really up to date about its AI knowledge the Vatican City like the Vatican has its own AI ethics framework The way for example Germany and the EU and the UK do too and and this is partly the work of the Franciscan order who Make sure that the pope is pope is kept up to date with new technological developments and AI is Particularly interesting for people in thinking from religious contexts because of the ways in which On the one hand you have questions of technology and human values But on the other hand you have questions about what it means to be human What it means to create things that are Made in the image of the human and particularly the Abrahamic religions Have some serious questions to ask about that that said Back when we did our research on the Middle East and North Africa we found that the conversation From the Islamic perspective it is is more was then much less developed. There have been some developments since But it What we did find is that it's not the case that having a certain religious perspective rules out thinking about or fantasizing about or Wondering about AI I mean again thinking from the Islamic perspective the Thousand and one nights are one of the story sets that have these ancient imaginings of artificial beings made of brass and copper in them And and other religions do all have have their own ways of Thinking about this. So actually the religious perspective is a really important one as a source for different ways of thinking about other intelligences Thank you so much any more from Slido That's it. Okay, then I have one last question maybe to wrap this up Conta if you allow me Maybe talk about AI. We also talk of course about hugely scalable amounts of data that are needed To for AI to process To produce those results and a large amount of data, of course that are produced while doing so and not only in Europe Of course, this is a also a huge question of data security Of how this data is stored by whom it is stored Is it large companies the GAFA companies or is it actually state-owned? Is it centralized? Is it deep decentralized now when we listen to you and your research now in the past 90 minutes we basically talked about a diversity of how we Should perceive even in the West how a is seen across the world, right? Which is sort of a movement of thought That would tend to decentralization, right? How we see AI has been too dominant to hegemonic in the West to put it plainly now What's your take now on the storage? Of course, we there's a there's those You know tech wars we might call it between the East and the West and Europe how to store this data And what to do with it? What's your take on that? Should we do the same thing with the idea of storing all this data? that you proposed actually in seeing AI so my take on on Decentralizing data storage, I think I mean on on data there are I Guess I guess two different perspectives and the one perspective is more data makes more AI and if the data set is currently biased that just means that Not enough people across the world have been added to the data set if the data set is biased in favor of White people that means that you just need to add more data of non-white people to the data set The other take is Well, that's that's all well and good But what are you going to use that data for and actually I'm quite alright with not being facially recognized Because the data was never trained on people like me I'm happy with that and especially People who are from groups that are systematically over surveilled already Systematically over-policed already might not want to be included in that data set Now I think that there are similar I guess Approaches and ways of thinking about How to imagine a future with intelligent machines on the one hand diversifying The the kinds of stories that we have here to make sure that the stories that come out of For example, California and Silicon Valley aren't biased but also some stories are just people's own and Don't make sense in different contexts and it's great to be Aware of them, but it doesn't necessarily mean that we can take them out Out of their Specific contexts. I think the same goes for for data that in some cases Improving the data set diversifying is great in some cases Keeping the data set small and local is also is also great. It depends on what it's going to be used for Thank you so much. So tomorrow it's down to Dalin to the Japanese German Center on a different part of town That we've been today here right at the spiree in Berlin Mitter. I think it is already here Thanks for this beautiful kickoff Being with us from England for making the journey and of course to you all people actually Coming into this live space. I enjoyed that very much. So have a good conference. Have a good evening. Thank you Country hall