AI 2041: Ten Visions for Our Future

The stories are terrific. Kai Fu used his extensive contacts and knowledge network as well as his own incredible depth and breadth of knowledge to think through a roadmap for what he expects AI capabilities will be 20 years from now. He bases his imagination-based forecasting on ongoing actual AI R&D developments, things currently being done by AI start-ups in China, US and other countries, and by his own very well-informed judgement on how to extrapolate today’s current state-of-best-industry-practice with deploying AI to what should realistically be possible at scale 20 years from now. Except - the book does not present this roadmap per say. Rather, the book conveys this information in a much more interesting, imaginative, and highly contextualized way. Based on Kai-Fu’s roadmap and undoubtedly based on intensive conversations and collaborative review with Kai Fu, the science fiction writer Chen Qiufan (aka Stanley Chan) creates 10 stories. Each story is a remarkably high-fidelity portrayal of real, every-day life in the year 2041. The use of AI capabilities at that future point in time is just part of how things happen, woven into the fabric of what everyday reality is, though 20 years from now. Each story revolves around capabilities enabled by a few key AI areas. In this sense, each story has a very focused set of themes it elaborates per AI capabilities. At the same time, the stories are not about the AI per say. Each story is deeply and realistically contextualized so that it is about selected aspects of human life, experience, and interaction, as well as the surrounding human social and political environment in the year 2041. The way each story is crafted, certain AI capabilities are deeply interwoven into what life is like and how things happen. Through the collection of stories, the reader can tangibly grasp and vividly experience how AI will very plausibly be changing the way we live and the nature of our lives and interactions across our personal and work-related dimensions. After each story, Kai Fu gives his analysis of the AI capabilities highlighted in that story. This part is non-fiction. Kai Fu gives his views on where certain AI capabilities will be (those highlighted or implied in that story), and why he thinks this is the case. This supporting analysis, written in an easily understandable and accessible way, is a terrifically informative complement to each of the seemingly very real science fiction stories crafted by Qiufan. The post-story analysis helps to reinforce key points on how the AI capabilities featured in the story are likely to evolve over the next two decades. This really is story telling at its best. It is the best effort I have come across to help one envision and nearly experience how aspects of everyday life and work and the broader social/political environment are likely to be 20 years from now because of the pervasive usage of increasingly capable AI . Yet, there are no assumptions of magic or miracles per future AI capabilities when they envision these human-centric stories set in a future setting. The projections of capabilities 20 years into the future are to a very large extent--- fact and reality based given what is already observable now across the spectrum of leading-edge research to state-of-the-industry-practice deployment. Of course, the two authors do allow themselves to do some very creative, though reality-grounded “imagineering” of this uncertain future, and this allows them to explore new possibilities of how people and society may function within the stories. The approach this book takes of combining “evidence-based” science fiction with supporting post-story fact-based analysis on AI technology directions helps me to envision the forthcoming AI-permeated future 20 years from now better than anything else I have come across. I suspect you will feel likewise after you read this book.

I took up this book as a friend of one of the authors and were impressed by both. With an unusual format of paired short stories (by Chen) and commentaries (by Lee), they introduced readers to AI technologies and their likely further developments in two decades, and more intriguingly, imagined how people would be living in such a profoundly AI-transformed world. Chen’s fictions are intricate, tender, and always with reassuring happy endings. His outlook is worldly although all stories show a touch of other-worldly almost fairy tale-like quality. Lee’s writing is excellent and authentic, imbued with authority, confidence, forward-looking and optimism, although the last of which is oddly (unintended, I’m sure) contrasted or undercut by an unmitigated melancholy tone of Chen’s tales despite happy endings. Overall, a very fun and interesting book. Congrats! However, I did get away with the impression that “the more it changes the more it remains the same”. As transformative and profound as the Authors believe AI technologies really are, in their imagined brave new world, we see racism, terrorism, social injustice, and human loneliness, desperation and deception – very much the same world we are living in right now. Can technology elevate a society at all? From this perspective, the book has provided but a limited vision.

I don’t read much sci-fi, so I didn’t know what to expect. However, when I finally dug in, I couldn’t put the book down. I have to say, if anyone out there is looking for the best sci-fi of 2021, this is exactly the kind of book you should be considering. (In fact I am not sure if you can even call this book sci-fi. Something more like a futurist approach to technorealism? Or an intriguingly fictionalized speculation of AI technology?) I have a background in artificial intelligence research (I majored in CS and philosophy back in college, and I currently work for a tech thinktank), and I am an avid lover of stories that take place in urban settings. This book pretty much encompasses all I want to read about the way AI interacts with our human society, and the writers did a GREAT job. The book comes in ten stories. Each story, happening exactly 20 years from now, takes place in a different city in the world (aka ten versions of the future in the year 2041). AI technology is at the heart of every story: deep learning, image recognition, GAN, deepfake…each story is paired with an essay explaining the central technology. My favorite stories are the one set in Africa and the one set in Korea. Though I would have liked to read about what becomes of globalization in 2041 and see more country-to-country interactions (climate change is one of the big topics the book tackles, what is the UN doing, for example?), but the writer Chen generally maintains a one-place-per-story approach, trying to include as many aspects of everyday life as possible. The themes include job seeking, COVID, data privacy, bitcoin—even virtual idols, can you believe it? Throughout the read, I am repeatedly reminded of The Age of Em by Robin Hanson — another work that imagines a historicized version of the future where robots roam the Earth that I picked up a few years ago. Hanson attempts to describe what a society would look like after “emulated minds” take over Earth, tackling the topic mostly through economics and psychology. However, I have to say, though AI2041 depicts a similar futuristic/robotcentric setting, it is flat-out better. Better in every aspect. And I think what made the crucial difference is the writing itself. The writer Chen Qiufan’s storytelling skills are superb. Although it says in the cover page that the book is a translation, I honestly don’t think any of the nuances had been lost in translation (the praise should also go to the translators). The stories read as fresh as new. The narrative flows smoothly. Chen seamlessly weaves together the technology and the stories and characters, and it does not seem forced at all. He gracefully engages his readers in each of the little “versions of the futures” taking place in different cities in the world. In love with Chen’s writing style, I looked up his name and found out that he has written another long form sci-fi, Waste Tide. Just ordered a copy, too! Then, the essays. To be honest, I was more excited about what it has to offer about AI, but the technology depicted in the book does not deviate so much from mainstream research and predictions. However, I can imagine that someone who wants to educate themselves on AI and its implications getting a lot out of their read. Lee’s essays are well-written and cuts straight to the point. He achieves clarity and precision without relying too much on the regular jargon, and I think this is another point of strength. I can even imagine this book being used in college as a kind of interdisciplinary, introductory textbook to artificial intelligence or science & literature. In general, I highly recommend this book. If you are someone like me, read it for the sake of the stories, at least!

Overview Like many people, I was caught off guard by how advanced chatbots had become when GPT-4 was released in March of this year. Indeed, it is now the number one topic of conversation in both tech and financial circles. Wanting to get better up to speed, I remembered this book, written by former Google employees involved in their AI efforts, which was recommended by Ray Dalio in his social media posts. The book approaches the subject primarily through presenting 10 short stories, written by Chen Qiufan. Each is preceded by a brief introduction by Kai-Fu Lee and is followed by a more detailed analysis also by Lee. The Short Stories The short stories cover many areas of AI such as: deep learning, natural language processing, augmented and virtual reality, along with autonomous vehicles and weapons. Quantum Computing, although having much broader applications than AI, is also discussed. Beyond just discussing the raw technology, the short stories also delve into the societal implications of AI: both its promises and its perils. Example themes include: what will losing jobs to AI feel like to those who lose them? How can AI work directly towards making humans happy and what must it avoid when doing that? In general, I spend little time reading fiction: only a classic here and there. I am, thus, far from a science fiction connoisseur. I am somewhat more open to sci-fi movies, but, still, only the real classics: the original three of the Star Wars series, Blade Runner, and the first couple of Terminators. Thus, I may not be the best person to comment on the quality of the science fiction. With these qualifications out the way, my thoughts on the short stories are: Although none of the short stories were terrible, I only found one, “Quantum Genocide”, to be highly engaging. The others run the whole spectrum from “quite boring” for “The Holy Driver” to “pretty interesting but somewhat cheesy” for “Contactless Love”. The former deals with autonomous vehicles, while the latter deals with AI in medical care. My broad observations were that when reading the short stories I did not find that they added much beyond their introduction and analysis by Lee. In many cases, I found myself bored, wanting to jump straight to the analysis, but feeling compelled not to miss anything. That being said, the trend is that the quality of the stories does improve the farther you get into the book. This includes becoming more interesting as the themes are of things on a grander scale. Overall, I would rate the short stories by themselves 3 out of 5. Unfortunately, there are many 2’s in that rating and only “Contactless Love” and “Quantum Genocide” rate a 4 and a 5, respectively. The Analyses Although it takes slogging through the short stories to get to it, I found Lee’s analysis to be much more engaging. He gives a decent introduction to the key AI technologies. Even more interestingly, I enjoyed his discussion of Quantum Computing and decided that is a subject I wanted to follow up on. None of the technologies are discussed at a deep technical level, there is certainly no math, but everything seems concise and accurate. Besides giving introductions to the technologies, Lee discusses, at length, the societal implications. Here he is particularly concerned with a coming onslaught of “deep fakes” (my phrase, not his); how society will deal with a situation in which many people will lose their jobs; and how some of the foundations of economic theory, such as scarcity, will be rattled to the core. In particular, Lee argues that the latter will necessitate a new social contract. In general, Lee does give time to both the dire and optimistic predictions. In some cases, he does not water down the implications. A prime example is his discussion of autonomous vehicles turning 'trolley problems' into reality. Here he presents it as what if your car has to choose between saving your child as a lone occupant and some number of random strangers. How should the car be programmed? Although clearly aware of the dire scenarios and generally giving at least some mention to them, Lee clearly leans toward thinking everything will all work out in the end. He is trying to be fair, by no means has his head in the sand, and is trying to present the dire scenarios. Despite this, I thought there were some big important, dire themes that he missed. For example, he discusses the promise of AI in education. In this case it seems like almost all roses. One thing Lee misses, however, is the potential AI has to be used for political indoctrination and to find out too much about children early on: For example, who is showing signs of being a budding leader of dissenting [insert ideology here]? In general, Lee seems to trust that AI is something that governments will make benevolent decisions about with, perhaps, only unintended consequences. There is some mention of social credit systems by Lee. Quifan also hints at the question of what happens to those who do not wish to participate in the new order in “The Holy Driver”. Overall the political dangers do not seem to discussed adequately, however. The Central Problem One reason I am not a big fan of science fiction is the recurring theme that humans possess qualities that robots and other species in the universe simply cannot replicate. In this book this manifests itself as Lee’s optimism that there will be jobs for humans since AI cannot match us in terms of creativity, strategic thinking and empathy. One of the most surprising things to many recently, however, is just how good AI has become in areas emphasizing creativity. This is particularly true of art, although its abilities in music are steadily gaining good reviews. It is quite surprising that GPT-4 seems to be even better in creative artistic endeavors which are “fuzzier” than in dealing creatively with more concrete mathematical problems. Similarly, Lee, to some degree, ends up arguing against his own point when he points out how AI has now become better than average humans at reading emotions from looking a pictures of people. When interacting with chatbot users some, including me, are starting to note that GPT is gaining the ability to feel your mood. For example, Bing, which is now powered by GPT-4, will now ask you to restart conversations if it feels you are being curt with it. If you are talking to it about work you feel proud of, it seems clearly programmed to be encouraging. Finally, Lee also does not spend much time discussing situations when AI blows up spectacularly other than to say it does not happen that often and we have to weigh that against all the times it is right. It was after this book was written, but the general optimism that some, like the authors, feel about AI might mean its security is not adequately analyzed. For instance, researchers recently discovered hacks to trick GPT into giving instructions for car jacking and bomb building. Overall I give Lee’s analysis a 4 out of 5. Summary The book seems to be targeted at those who enjoy science fiction and who are seeking a general introduction to the technology behind AI and its societal implications. It is written by authors who clearly believe that the promise of AI outweighs its perils. It is good to hear their perspective. The authors do not have their heads in the sand. Most of the book, indeed, discusses the perils of AI. Despite this, there were enough omissions that I felt the perils were still not adequately discussed. Finally, I did not find that the fiction added much beyond what was discussed in the analyses. Some, but not all, of it was boring. Although the book’s manner of presentation might work for others, I would have preferred some of the space devoted to the fiction to have been spent exploring the perils of AI even more deeply instead.

I really enjoyed AI 2041: Ten Visions for Our Future and have already recommended it to several people. It’s one of those books that manages to be both engaging and informative without feeling overly technical or dry. The format works especially well, each chapter blends a fictional story with real-world analysis, which makes complex AI concepts easier to understand and actually interesting to think about. Instead of just explaining the technology, it shows how it could realistically shape everyday life in the near future. What stood out to me is how balanced it feels. It doesn’t lean too heavily into hype or fear. Instead, it presents a range of possibilities...some exciting, some uncomfortable...and lets you think through the implications yourself. It’s a great read if you want to stay informed about where AI is heading without having to dig through dense, academic material. Definitely a solid addition to any “informed reading” list, but also enjoyable enough to recommend just as a good book on its own.

Unique and creative thoughts in exploring how the future and current technologies interact with a positive outlook rather than fears.

To say that AI 2041 is enlightening and valuable, is to understate its significance. It’s organized around ten short stories - vignettes of future life transformed by advances in AI and data science that cover how we will learn, amuse ourselves, live in cities, deal with pandemics and more. Each story is sandwiched between a non-fiction introduction and an epilogue that enable the reader to relate the likely trajectory of technology’s evolution to its impact on our social fabric. In so doing, the authors have created a new genre of “scientific fiction” that is sure to enlighten anyone who wonders where the heck this is all heading. Underneath its stories, AI2041 poses a more fundamental question that can be summed up as “so what for the humans?” Specific ponder-worthy topics include: “What is a career?” “How is our sense of what it means to be human likely to evolve?” “What will being ‘educated’ mean in the future?” “How will humans and machines come together in new forms of hybrid intelligence and what will be the new rules of engagement?” Much has been made of how AI will encroach on the “trade space” for human labor. As AI devours jobs involving routine thinking and pattern recognition, we will begin to bid farewell to countless professional activities such as reading x-rays, legal research and basic accounting. We can be certain that the pace of innovation will continue to accelerate as capital continues to pour into “AI everywhere” business models and their associated ventures. It is also inevitable that we will need to continuously ‘negotiate’ our relationship with AI as it becomes ever more capable. It’s worth noting that back in the 1990’s, author Dan Simmons described a future in which AI would not only coexist with humans but would eventually declare its independence and develop multiple competing embodiments and relationships with humanity while it pursued its own creative project of developing an “ultimate intelligence.” So stay tuned. One of the key contributions of AI2041 is to show us how to thrive in a world increasingly shaped by technology. In a recent interview with me, Lee Kai-fu referred to the secret sauce of humans in the AI era as “warm skills.” I think this is a fundamental insight that calls out for more clarity about which proficiencies are inherently human and can never be replaced by technology. These include empathy, compassion, collaboration, a growth mindset, agility, trust building, and creativity among others – what some refer to as “21st century skills.” In the vintage short story Virtuoso by Herbert Goldstone, a robot asks an eminent pianist for permission to learn the piano. By day’s end, it plays Beethoven's Appassionata sonata with such feeling that it brings tears to the maestro’s eyes. The story concludes with the robot refusing further involvement music, saying that “some things were not meant for machines.” The point here is that while a robot may be able to perform music perfectly, at the end of the day it doesn’t really matter. Music is an inherently human activity, requiring human empathy and aesthetic sense to make a meaningful connection between performer and listener. What does matter is when humans create the work, perform it, appreciate it and are moved by it. And when we pay attention to the humanity in a performance – including its imperfections and idiosyncrasies – we express the kind of warm skills that will forever define our human ‘trade space.’ Science fiction has historically opened our eyes to far horizons. AI2041’s scientific fiction gives us a way to open our eyes to what is actually going on all around us and where things are heading.

A very complete view of the probable future with a lot of detail. I recommend to read the chapter analyses first.

Very cool book, loved the quality of the people-focused short stories. The book has a real global flavour to it. Each short story is followed by its own "Now for the science..." bit. Very consumable, enjoyable and informative. For a topic that can be very heavy technically, this book does a really good job of making it more about what AI might do for people.

This is a long book for what it purportedly sets out to do, but it is enjoyable. Each of the ten stories is well written. While not sophisticated stories, they have enough depth to be enjoyable while delivering the vision of the future. Each of the stories is followed by an explanation of the various technologies and how they work (at an overview level) and also the non-technical aspects such as bias, economics, social impact, etc. If you’re interested in the wide world of the AI, this would be a very interesting read for you.

Great book with many interesting insights about AI. I highly recommend!

Incredibly interesting, paints a possible view of the future and also provides in-depth technical analysis

This story actually has multiple small stories how AI could be like in our lives. The good, and the bad.😐