AI 2041: Ten Visions for Our Future

This entry covers “AI 2041: Ten Visions for Our Future” (Crown Currency, September 2021) as one portrait of the year 2041, assembled from Chen Qiufan’s ten short stories and the analysis Kai-Fu Lee writes after each, and not the stories as separate works. Lee calls the book “a thought experiment describing how AI will transform the world in twenty years”: he forecast which technologies would be mature by 2041, Chen set one story in each of ten countries inside that forecast, and Lee follows each with an essay on the technology and its consequences. The premise comes from Lee’s introduction. Deep learning succeeded because decades-old algorithms met abundant data and cheap computation; it keeps improving on “single-domain tasks” and will beat people on more of them every year, but general intelligence would need “a dozen more breakthroughs like deep learning” after only one in sixty years, so the book assumes no AGI by 2041 and leaves creativity, strategy, reasoning, and emotion to humans. Book site; publisher; Time excerpt; ITU interview

The transition is twenty years of curves rather than an event. Lee’s rule is that people overestimate what a technology does in five years and underestimate what it does in twenty, so each story opens on a world already changed and its analysis names the curve. In Mumbai a deep-learning insurer prices a family from everything it can observe, nudges their behavior, and learns caste prejudice from its data (The Golden Elephant). In Lagos deepfakes are as easy as a photo filter and detection always lags (Gods Behind the Masks). In Korea orphaned twins are raised by AI tutors that shape them differently (Twin Sparrows). In Shanghai a pandemic-scarred woman lives among household robots, autonomous delivery, and wearables (Contactless Love). In Tokyo a fan investigates a virtual idol’s death in extended reality (My Haunting Idol). In Sri Lanka a gamer learns he is a remote “ghost driver” for autonomous vehicles in the emergencies they cannot handle (The Holy Driver). In Iceland a grieving physicist uses quantum decryption and autonomous drone swarms to assassinate world leaders, and Lee calls autonomous weapons the third revolution in warfare (Quantum Genocide). On an island near Qatar an AI optimizes a happiness metric for wealthy guests and fails, because the metric is not happiness (Isle of Happiness). Story notes; Diamandis summary

Two stories carry the economic mechanism. In The Job Savior, set in the United States, automation accelerated by the pandemic strips out routine work, the government answers with a universal basic income, and the stipend fails: addiction, gambling, crime, and suicide rise among people with money and nothing to do, and the program is repealed within about a decade. A “job reallocation” industry replaces it. Synchia, where the protagonist Jennifer Greenwood trains, contracts with employers such as the automating Landmark Construction to retrain displaced workers with some dignity; its rival OmegaAlliance claims a perfect reassignment rate by fitting workers with VR headsets to perform simulated tasks of no real value. Lee’s analysis says the explicit goal of AI is to take over human tasks, estimates that 30 to 40 percent of jobs are replaceable within twenty years, expects human-to-human service work to grow as populations age, and offers the “3 Rs,” relearn, recalibrate, renaissance, while conceding that income without purpose is what sank UBI.

In Dreaming of Plenitude, set in Australia, the government’s Project Jukurrpa issues a Basic Life Card that covers food, water, clothing, shelter, and energy for free, and a second currency, Moola, that is earned through community service and reputation and steers people toward care, education, and creative work. Keira, a caregiver, is paired with Joanna, an aging marine scientist in the early stages of Alzheimer’s, in a smart retirement community. The Moola system is gamed by people chasing status. Lee’s analysis makes clean energy near 3 cents per kilowatt-hour, automation, and additive manufacturing the drivers pushing necessities toward free; he prefers “plenitude” to post-scarcity, expects it to start with necessities and widen, and says that once work is optional the open question is what motivates people. Noema on plenitude

Humans keep control by three means: regulation, which Lee says will always lag innovation and is still necessary; technical fixes such as authenticating content at capture, since he believes most AI harms are solvable by technology; and a division of labor in which AI does quantitative and routine work while people supply creativity, strategy, and care, with humans as backup in edge cases. His stance is that AI “will not love you back,” that it will do more good than harm, and that humankind remains the author of its destiny.

For the library, the mechanism is diffusion: narrow deep learning spreads through insurers, schools, cities, employers, and governments over twenty years, with no takeover and no single turning point. The book is specific about each domain’s technology and side effects, but the transition is asserted from curves and probabilities rather than shown; Lee names no year for the driving jobs he expects to vanish, and the UBI failure is a story premise, not an argued result. Left open are governance and geopolitics, which the stories barely touch; whether displacement is absorbed, since The Job Savior gives no answer to the purpose question; and whether plenitude arrives, which depends on clean energy keeping pace with computing. Reviewers find the authors light on downsides and the book without a conclusion. Rate specificity on the per-domain mechanisms and Lee’s stated curves, plausibility on the no-AGI premise, and the outcome scales on the assembled 2041. It sits beside What 2026 Looks Like as a second capability-curve portrait of ordinary life, and beside Machines of Loving Grace as a second optimistic portrait, this one without powerful AI. KPMG Compact review; brokensandals review