At the 2000 Stanford conference, Kurzweil was the only person who felt AGI would happen by 2029, 30 years out.

Judged as of the event Kurzweil describes, a Stanford conference around 1999–2000 on AI timelines (the 'Spiritual Robots' symposium or a similar event). True if contemporaneous records show that Kurzweil predicted human-level AI or AGI by 2029, roughly 30 years out, and that no other participant gave an estimate that early. Records could include transcripts, reports, a poll of attendees or participants' accounts. Mostly true if the substance holds but a detail is off: the year of the conference is slightly different, or the 'roughly 30 years' framing does not quite fit. Mixed if Kurzweil made the 2029 prediction, but the records are split on whether others gave similar or earlier timelines. Mostly false if the conference took place and Kurzweil gave an early timeline, but at least one other participant also expected AGI within about 30 years. False if records show that he did not predict 2029 there, that no such conference took place, or that several participants held the same view.

Set before the research began.

1. E1 The Last 6 Decades of AI — And What Comes Next: Ray Kurzweil (Transcript) SingJupost (transcript of Ray Kurzweil's TED 2024 talk) · 8 July 2024 Own record

   “And hundreds of AI scientists came from around the world. And they agreed that it would happen. We would achieve AGI, but not within 30 years. The estimate was 100 years.”

2. E2 The Last 6 Decades of AI — And What Comes Next: Ray Kurzweil (Transcript) SingJupost (transcript of Ray Kurzweil's TED 2024 talk) · 8 July 2024 Own record

   “So I estimated 2029. That was met with a lot of skepticism. Stanford had actually been monitoring my predictions. They called an international conference to talk about my prediction.”

3. E3 Ray Kurzweil: A.I. Is Not Hollywood's Alien Invasion Observer · 18 September 2025 Own record

   “Stanford University was so concerned that they organized an international conference to assess my prediction. Several hundred A.I. experts from around the world came, and 80 percent agreed that a computer would match human performance eventually, but most thought it would take 100 years.”

4. E4 AI prediction case study 4: Kurzweil's spiritual machines — LessWrong LessWrong (post by Stuart Armstrong, co-authored with Kaj Sotala and Seán ÓhÉigeartaigh) · 14 March 2013

   “and made several predictions for the years 2009, 2019, 2029 and 2099.”

5. E5 AI prediction case study 4: Kurzweil's spiritual machines — LessWrong LessWrong (post by Stuart_Armstrong) · 14 March 2013

   “made several predictions for the years 2009, 2019, 2029 and 2099.”

6. E6 Predictions made by Raymond Kurzweil Wikipedia

   “Kurzweil also believes a machine will pass the Turing test by 2029.”

7. E7 AI prediction case study 4: Kurzweil's spiritual machines — LessWrong LessWrong (comment by CarlShulman) · 14 March 2013

   “he predicts that Turing-Test passing AI will be around by 2029, well before he predicts brain emulation”

8. E8 Will Spiritual Robots Replace Humanity by 2100? organism.earth · 1 April 2000

   “So I see that this is the beginning of a long evolution—several decades at least, about the same time scale as the speakers have all been talking about; say, thirty years, thirty, forty years even.”

The only accounts in the evidence of Kurzweil being alone in predicting 2029 at the Stanford conference are his own later retellings (E1, E2, E3). The subject restating his own claim does not count as independent evidence. Independent sources (E4–E7) confirm that Kurzweil predicted 2029 for Turing-test-level AI, but they say nothing about the conference or what other participants expected. E8, a contemporaneous record of the April 2000 Spiritual Robots symposium, has one speaker saying the other speakers had been talking about a time scale of roughly 30 to 40 years. That hints others gave similar timelines, but the quote does not say whether that time scale refers to human-level AI or AGI, or who gave which estimate, so the evidence does not settle the claim either way.