Both Turing Award laureates; pioneers of symbolic AI and cognitive simulation

“Within ten years a digital computer will be the world's chess champion, unless the rules bar it from competition.”

Why incorrectIn 1968 the strongest chess programs were rated well below master level; a computer did not take the world title until Deep Blue in 1997.

Turing Award laureate and Nobel laureate in Economics; a founding figure of AI

“Machines will be capable, within twenty years, of doing any work a man can do.”

Why incorrectBy 1985 no general-purpose machine could perform arbitrary human work; symbolic AI had stalled and the second AI winter was underway.

Bletchley Park cryptanalyst; originated the intelligence-explosion argument

“It is more probable than not that, within the twentieth century, an ultraintelligent machine will be built.”

Why incorrectThe twentieth century closed with no machine matching, let alone exceeding, general human intelligence.

Turing Award laureate; co-founder of the MIT AI Lab

“Within a generation the problem of creating 'artificial intelligence' will be substantially solved.”

Why incorrectBy 1997 AI remained a collection of narrow techniques; the general-intelligence problem was not substantially solved and is not solved today.

Well-known chess master who publicly wagered against strong computer chess

“I am prepared to wager £1,250 that no computer program will beat me in a match within ten years.”

Why incorrectLevy won the original 1978 wager, but subsequent bets he extended were lost as computer chess advanced far faster than he expected; by 1989 Deep Thought defeated him convincingly.

Turing Award laureate; co-founder of the MIT AI Lab

“In from three to eight years we will have a machine with the general intelligence of an average human being.”

Why incorrectNo general-intelligence system existed by 1978; the underlying symbolic programme did not scale as expected.

Robotics pioneer; namesake of Moravec's paradox on sensorimotor difficulty

“I am confident that this exponential growth will lead to computers with humanlike abilities by around 2010.”

Why incorrect2010 arrived with impressive narrow AI but no system exhibiting humanlike general capability; the hardware extrapolation held, the capability did not.

Influential futurist and inventor; author of *The Age of Intelligent Machines*

“A computer will defeat the world chess champion by 1998.”

Why incorrectThe chess claim was correct; Kurzweil packaged it with 1990s claims — speech-to-text replacing keyboards, real-time phone translation — that did not arrive on schedule.

Science-fiction author and mathematician who popularised the modern term 'Singularity'

“Within thirty years, we will have the technological means to create superhuman intelligence. Shortly after, the human era will be ended.”

Why incorrectThe thirty-year window closed in 2023 without superhuman intelligence and without the end of the human era.

Co-founder of MIRI; among the most influential early writers on AI existential risk

“Computing speed doubles every two subjective years of work. Two years after Artificial Intelligences reach human equivalence, their speed doubles. One year later, their speed doubles again. Six months — three months — 1.5 months … Singularity.”

Why incorrect2021 arrived with no Singularity, no human-equivalent AI capable of recursive self-improvement, and no runaway speed explosion.

Influential futurist and inventor; author of *The Age of Spiritual Machines* and *The Singularity Is Near*

“A $1,000 computing device is now approximately equal to the computational ability of the human brain… Computers routinely pass the Turing Test.”

Why incorrectIn 2019 no system routinely passed the Turing Test; ubiquitous VR embodiment and machines claiming consciousness did not occur.

Co-founder of MIRI; among the most influential early writers on AI existential risk and rationality

“I believe that there is a greater than 70% chance that molecular nanotechnology will be developed within the years 2003–2015… nanotechnology, by default, will end up killing everyone.”

Why incorrectNo transformative molecular nanotechnology was developed in the 2003–2015 window, and none of the extinction scenarios he described occurred.

Co-founder of what became MIRI; a foundational writer on the alignment problem

“I would guess that our final stage of seed AI, going critical, will occur between 2005 and 2020, probably around 2008 or 2010.”

Why incorrectNo 'seed AI going critical' occurred by 2010, or by 2020; the recursive-self-improvement pathway he described was not realised.

Influential futurist and inventor; author of *The Singularity Is Near*

“By 2020 a $1,000 personal computer will have the processing power of the human brain… and by 2029, a computer will pass the Turing Test.”

Why incorrectThe practical claim — a $1,000 machine running a human-equivalent mind — did not occur by 2020 on any operational benchmark.

Founder of Palm Computing and Numenta; neuroscience-inspired AI theorist

“I believe we will build genuinely intelligent machines within the next few decades… I would be surprised if it took more than fifteen years.”

Why incorrectBy 2020 no system met Hawkins's own memory-prediction framework for genuine intelligence.

DeepMind co-founder and Chief AGI Scientist; co-author of one of the standard formal definitions of intelligence

“I think human level AI will be passed in the mid 2020s… my modal estimate would be around 2025, with a distribution that puts significant mass on the years around 2018.”

Why incorrectThe lower tail he flagged as plausible (around 2018) passed without human-level AI on his own criteria.

Theoretical physicist and best-selling popular-science author

“By 2020, we will have robots as intelligent as a mouse. By 2030, as intelligent as a dog.”

Why incorrectMouse-level embodied cognition — sensorimotor generality, navigation, self-directed learning — was not achieved by 2020.

Co-founder of Sun Microsystems; prominent Silicon Valley venture capitalist

“Eighty percent of what doctors do will be replaced by technology — machine learning, algorithms, and machines — within the next decade.”

Why incorrectA decade later physician employment had grown, not fallen by 80 percent; machine-learning tools were adopted as narrow assistive aids rather than as replacements for physician work.

CEO of Tesla, SpaceX and xAI

“Tesla will have full self-driving cars, at a reliability level that we would consider no hands on the wheel, in a five-year time frame.”

Why incorrectFive years later Tesla's driver-assistance systems still required constant active human supervision and were not certified for hands-off operation.

Led Google's self-driving programme (later Waymo); co-founder of Aurora

“My oldest son is 11, and he'll be 16 in four and a half years… my team is committed to making sure that doesn't happen — that he doesn't need a driver's licence.”

Why incorrectBy 2020 no consumer self-driving service let a teenager forgo a driver's licence; Waymo's driverless rides remained a limited service in parts of Phoenix.

One of the 'godfathers of deep learning' and a Turing Award laureate

“People should stop training radiologists now. It's just completely obvious that within five years deep learning is going to do better than radiologists.”

Why incorrectAs of the mid-2020s the number of practising radiologists in the US has grown, not shrunk; the profession remains in shortage. AI is used as a workflow aid in radiology but has not replaced radiologists.

Co-founder of Google Brain and Coursera; former Baidu Chief Scientist

“If a typical person can do a mental task with less than one second of thought, we can probably automate it using AI either now or in the near future.”

Why incorrectFive years on, automating one-second mental tasks in production remained narrow and brittle across most real-world domains.

Founder and CEO of Meta (formerly Facebook)

“One of our goals for the next five to ten years is to basically get better than human level at all of the primary human senses: vision, hearing, language, general cognition.”

Why incorrectA decade on, systems match or exceed humans on narrow benchmarks but not across vision, hearing, language, and general cognition as a whole; general cognition in particular remains open.

Founder and CEO of SoftBank Group; major AI investor

“In the next 30 years, the singularity will happen… Within one year, computers will surpass human beings on the IQ test.”

Why incorrectNo general-purpose system surpassed humans on standard IQ tests within a year of the keynote.

Former president of Google China; founding director of Microsoft Research Asia; venture capitalist and AI author

“AI will be able to replace about 40 percent of the world's jobs within 15 years.”

CEO of Tesla, SpaceX and xAI

“We are headed toward a situation where AI is vastly smarter than humans. I think that time frame is less than five years from now.”

Why incorrectBy 2025 frontier systems exceeded humans on many narrow benchmarks but did not meet common definitions of general intelligence surpassing any single human across arbitrary tasks.

CEO of OpenAI

“In the next five years, computer programs that can think will read legal documents and give medical advice. In the next decade, they will do assembly-line work and maybe even become companions.”

Why incorrectFive years on, no AI system reliably 'reads legal documents and gives medical advice' at professional standard without human oversight; deployment remains assistive rather than autonomous, and no system does assembly-line work or serves as a general companion.

Co-founder of DeepMind and Inflection AI; CEO of Microsoft AI

“The Modern Turing Test: an AI that can be given a $100,000 investment and turn it into $1 million within a few months, largely on its own. I think that's coming in the next five years.”

Co-founder and CEO of Google DeepMind; 2024 Nobel laureate in Chemistry for AlphaFold

“I think we're some way away from AGI… I would say maybe in the next decade, so a five- to ten-year kind of time frame.”

CEO of Tesla, SpaceX and xAI

“AI will probably be smarter than any single human next year. By 2029, AI is probably smarter than all humans combined.”

Why incorrectThe 'next year' target (2025) has passed without AI systems demonstrating general intelligence surpassing any single human across the full range of cognitive tasks.

Best-selling public intellectual widely cited on AI, politics, and long-run history

“It is quite likely that in the next presidential election in the United States, in 2028, AIs will already play a bigger role than humans.”

Co-founder and CEO of Anthropic; former VP of Research at OpenAI

“AI could wipe out half of all entry-level white-collar jobs — and spike unemployment to 10 to 20 percent in the next one to five years.”

Lead author is a former OpenAI governance researcher; the scenario has been widely discussed as a concrete short-timeline forecast

“By early 2027, AI systems are automating AI research itself; by the end of 2027, superintelligence has arrived.”

CEO of OpenAI

“We are past the event horizon; the takeoff has started. Humanity is close to building digital superintelligence… by 2030, intelligence and energy will be wildly abundant.”

Influential NYT tech columnist and co-host of the Hard Fork podcast

“I believe that very soon — probably in 2026 or 2027, but possibly as soon as this year — one or more A.I. companies will claim they've created an artificial general intelligence. But even if A.G.I. arrives a decade later than I expect — in 2036, rather than 2026 — I believe we should start preparing for it now.”

Co-founder of DeepMind and Inflection AI; CEO of Microsoft AI

“I think that in the next 12 to 18 months, most white-collar tasks are going to be fully automated.”