AGI Zeitgeist

Panel evidence

A13 panel checks: group d2

A working document. It is one of the documents the site is built from, published as it is written for anyone checking the work, and it uses the project's own shorthand. The plain account of the same material is on the home page.

A13 panel checks: group d2

Ground rules followed: no edits under the project folder; every item below comes from a page I fetched or a search result that showed it, and is marked as such. BBC, Guardian, Independent, Telegraph, New Scientist, Wired and FT pages cannot be fetched or searched by domain from this environment, which is why most of the missing outlet coverage (C3) is missing. Treat "not found" as "not found from here", not "doesn't exist".

Murray Shanahan

Tier checked: core C1 earliest: 2014-07 (TechRadar interview; KitGuru write-up dated 2014-07-09) | timing | "So within the next 30 years I would say better than fifty-fity that we'll achieve human-level AI." (typo as printed) | https://www.kitguru.net/gaming/anton-shilov/ai-will-match-human-intelligence-in-the-next-30-years-robot-expert/ (fetched; the original TechRadar page https://www.techradar.com/news/world-of-tech/ai-will-likely-match-human-intelligence-in-the-next-30-years-says-robot-expert-1256532 only loaded a paywall/header) | secondhand (a direct interview quote, passed on by a second site)

Ben Goertzel

Tier checked: core C1 earliest: 2006 (talk "Ten Years to a Positive Singularity (If We Really, Really Try)", TransVision 2006, Helsinki, given by video; date and venue from a search result) | timing | "If we made a positive Singularity a real focus of our society, I think a ten year time-frame or less would be eminently possible." | https://www.thekurzweillibrary.com/we-could-get-to-the-singularity-in-ten-years (fetched; republished 2014-12-26) | primary (his own essay, reprinted). Backed up independently by the New Atlantis, 2009-10-03, https://thenewatlantis.com/?p=4941 (fetched; secondhand).

Michael Wooldridge

Tier checked: core C1 earliest: 2017-10-10 (oral evidence to the House of Lords Select Committee on AI) | possibility | "the Hollywood dream of conscious machines is not imminent, and indeed I see no path taking us there." | https://cs.ox.ac.uk/innovation/research-impact/case-AI.html (fetched; Oxford's in-house Inspired Research newsletter, Winter 2017) | secondhand (his department's write-up quoting his testimony; the committee transcript would be the primary source but I didn't fetch it)

Judea Pearl

Tier checked: core C1 earliest: 2011-12-27 (JTA, the Jewish Telegraphic Agency news service, by Tom Tugend) | possibility | "everything man can do, robots can do better" | https://jta.org/?p=442604 (fetched) | secondhand. CAUTION: the same story as syndicated in Intermountain Jewish News, 2011-12-29 (https://www.ijn.com/judea-pearls-futuristic-quest-the-perfect-robot/, fetched), presents this as reported speech: Pearl "believes that, at least in theory, everything man can do, robots can do better". It may be a paraphrase, not his words. A verbatim line from the same piece, less directly about possibility: "The only way to learn more about ourselves is by programming robots to emulate our behavior". Lead: his "Robots and Free Will" video interview, 2011-07-21 (listed on https://bayes.cs.ucla.edu/stat_bio.html, fetched; video not watched). C2 latest: 2024-09-09 | "I believe this will be in my lifetime, even though I'm not young." (about superintelligence) | https://en.globes.co.il/en/article-1001488789 (fetched; Globes, Israeli business daily) | primary (interview quote) C3 pieces: 1. 2011-12-27 | JTA (a news agency, but one covering Jewish news, so it may not count as "national") | "Judea Pearl, father of slain WSJ reporter, is a leader in artificial intelligence" | https://jta.org/?p=442604 | fetched. Partly about AI's future (robots outsmarting humans). 2. none found 3. none found. (His Turing Award coverage in 2012 is about his past work, not where AI is going.) Quota tags: philosopher: NEAR MISS. His degrees are in electrical engineering and physics (https://bayes.cs.ucla.edu/stat_bio.html), and he spoke at the 2011 Cognitive Science Society meeting as winner of its Rumelhart Prize (same page), but he has no philosophy or cognitive science training or appointment. Leave off. | skeptic: no (2011: "at least in theory" robots can do everything humans can; no doubt about timing found) | non_us_uk: no (UCLA since 1970) | frontier_lab: no Verdict: fails. C1 is thin (a possible paraphrase in a specialist news agency), C2 passes, C3 has at most one arguable piece. Most of his public commentary on human-level AI starts with The Book of Why (2018) and the 2018 Quanta interview. Queries: (1) Judea Pearl "human-level" OR "strong AI" OR "general AI" possible interview 2012…2017; (2) Judea Pearl Edge 2015 "What do you think about machines that think" response (he is not on the contributor list); (3) Judea Pearl Turing Award 2012 interview robots "free will" OR "consciousness" machines will; (4) "Judea Pearl" "Robots and free will" interview; (5) Judea Pearl future of artificial intelligence 2012…2016 Wired OR "Scientific American" OR "New Scientist" OR "Technology Review" OR "New York Times" robots think. (5 of 8 used.)

Max Tegmark

Tier checked: core C1 earliest: 2014-05-01 (The Independent op-ed, co-signed by Hawking, Russell, Tegmark and Wilczek) | possibility | "Success in creating AI would be the biggest event in human history." (I saw this line in a search result; the fetched copy confirms the op-ed, the date and that the piece argues highly intelligent machines are not "mere science fiction") | original http://www.independent.co.uk/news/science/stephen-hawking-transcendence-looks-at-the-implications-of-artificial-intelligence--but-are-we-taking-ai-seriously-enough-9313474.html (Independent can't be fetched from here); copy fetched at https://www.geneticsandsociety.org/article/stephen-hawking-transcendence-looks-implications-artificial-intelligence-are-we-taking-ai | primary, but co-written by four people, so the words can't be pinned on Tegmark alone. Earlier lead: his book Our Mathematical Universe (January 2014), not checked. C2 latest: 2025-06-03 (Epoch TV interview; reposted 2025-09-07) | "we're much closer to figuring out how to build this stuff than we are figuring out how to control it." | https://lifeboat.com/blog/2025/09/max-tegmark-can-we-prevent-ai-superintelligence-from-controlling-us (fetched) | secondhand repost of a primary interview. Note: this implies nearness without a date. The more explicit "within three years" claim (December 2023, seen in search results only) is earlier than the 2024 cutoff. A January 2026 Wall Street Journal profile (seen in search result) wasn't checked. C3 pieces: 1. 2014-05-01 | The Independent | "Transcendence looks at the implications of artificial intelligence - but are we taking AI seriously enough?" | URL above | fetched via a copy (the original domain is blocked). Caveat: this is his own co-written op-ed, which may fall under the rule excluding the person's own outlets. 2. 2015-01-16 | Christian Science Monitor | "Elon Musk spends $10 million to stop robot uprising" | https://csmonitor.com/Science/2015/0116/Elon-Musk-spends-10-million-to-stop-robot-uprising | fetched (names "MIT professor and FLI founder Max Tegmark"; FLI is the Future of Life Institute) 3. 2017-08-29 | Wall Street Journal | "When Machines Run Amok" (Frank Rose's review of Life 3.0) | reprint at https://www.frankrose.com/when-machines-run-amok.html | fetched (the reviewer's own reprint; the WSJ URL isn't given) Backup if piece 1 is disallowed: the Monitor piece quotes Tegmark speaking to The Atlantic (January 2015, Atlantic article not fetched). Quota tags: none apply. skeptic: no (warning of transformative AI from 2014) | non_us_uk: no (Swedish-born, but at MIT, United States, for all his pre-2017 statements) | philosopher: no (physicist) | frontier_lab: no (he co-founded the Future of Life Institute, a nonprofit, not a lab building AI models) Verdict: passes, with two flags: C1 is a co-written op-ed, and the C2 quote implies nearness without a date. Easy to strengthen C2 from his October 2025 Statement on Superintelligence work if needed. Queries: (1) Hawking Tegmark Russell Wilczek 2014 "Transcendence" "taking AI seriously enough" op-ed; (2) Max Tegmark 2025 AGI "within" years superintelligence said interview; (3) Max Tegmark 2026 superintelligence AGI timeline statement; (4) "Max Tegmark" Future of Life Institute Musk $10 million 2015 artificial intelligence article; (5) "Tegmark" artificial intelligence superintelligence 2016 OR 2017 "Life 3.0" review OR interview Washington Post OR "New York Times" OR "Scientific American" OR Wired. (5 of 8 used.)

Melanie Mitchell

Tier checked: core (C1, C2 and quota tags only; another helper is doing C3) C1 earliest: NOT FOUND from 2017 or earlier (looked again as asked). Nothing in search results ties Complexity: A Guided Tour (2009) to a statement on whether or when human-level AI arrives; I couldn't read the book text. The earliest dated items found are from 2018 or later: Complexity Digest episode 2018-03-09 (https://comdig.cssociety.org/2018/03/09/exploring-artificial-intelligence-with-melanie-mitchell/, fetched; no quote on the page itself) and a Singularity Weblog interview 2020-09-23 covering "the Technological Singularity and its possible timeline" (https://www.singularityweblog.com/melanie-mitchell/, fetched; no quote shown). Her widely cited New York Times op-ed of November 2018 also falls after the cutoff. Effort: 3 searches (the tool turned the third into 4), plus about 6 fetches. Next step if needed: the full text of Complexity (2009), chapters on AI and Hofstadter. C2 latest: 2025-09-15 | "Friedman states (without evidence) that 'the consensus within the A.I. community' is that superintelligent AI is coming, likely soon." | https://aiguide.substack.com/p/magical-thinking-on-ai (fetched; her own newsletter) | primary. A later post, 2026-06-09 (https://aiguide.substack.com/p/on-ai-and-jagged-intelligence, fetched), only points to her Yale Review essay with no AGI line. A 2024-10-16 video page (https://lifeboat.com/blog/2024/10/ai-is-intelligent-but-that-doesnt-mean-its-smart-dr-melanie-mitchell, fetched) has a paraphrase only. Quota tags: skeptic: CANNOT TAG. Her current reputation is skeptical, but the rule needs pre-2017 statements and I found none. | philosopher: LIKELY (she works as a cognitive scientist: analogy-making, the Copycat program, PhD under Douglas Hofstadter), but I fetched no dated source for her training or appointment in this pass, so it needs a source before it is applied. | non_us_uk: no (Portland State and Santa Fe Institute, United States) | frontier_lab: no Verdict: fails C1 as things stand (no statement from 2017 or earlier found); C2 passes. Queries: (1) "Melanie Mitchell" Portland State human-level AI OR "general intelligence" OR singularity 2014…2017 interview; (2) "Melanie Mitchell" "Complexity: A Guided Tour" computers human-level intelligence quote OR Kurzweil singularity 2009; (3) "Melanie Mitchell" said artificial intelligence "long way" OR "far from" human 2016 OR 2017 AlphaGo OR "deep learning" (the tool ran four searches for this one, so the budget of 4 was exceeded).

Oren Etzioni

Tier checked: core (C1, C2 and quota tags only) C1 earliest: 2014-12-16 (New York Times, John Markoff, "Paul Allen Adds Oomph to AI Pursuit") | possibility/timing | "we have gotten no closer to the larger goal of true human-level systems" (reported speech, not a direct quote) | http://www.nytimes.com/2014/12/16/science/paul-allen-adds-oomph-to-ai-pursuit.html (via University of Washington computer science news, https://news.cs.washington.edu/2014/12/15/allen-institute-for-artificial-intelligence-in-ny-times, fetched) | secondhand

Ernest Davis

Tier checked: core (C1, C2 and quota tags only) C1 earliest: 2014-10-23 (preprint date; published in Artificial Intelligence vol. 220, March 2015, pp. 121–124) | timing | "(Fortunately, it would not be necessary for such restrictions to have any impact on AI research and development any time in the foreseeable future.)" | https://cs.nyu.edu/~ed1/papers/Bostrom.pdf (fetched and read in full) | primary. "Such restrictions" means limits that would "halt AI research far short of human intelligence", so the sentence says that human-level AI is not in the foreseeable future. Also: "Certainly a general artificial intelligence is potentially dangerous; and once we get anywhere close to it, we should use common sense". C2 latest: 2025-07-22 (with Gary Marcus) | "Eric Schmidt's recent prediction...that there will be AI-based 'world-class mathematicians' within a year, is pure fantasy." | https://garymarcus.substack.com/p/deepmind-and-openai-achieve-imo-gold (fetched) | primary, co-written. Closer to AGI itself: 2025-01-05 (with Marcus): "It is only once AI researchers grapple directly with commonsense reasoning...that the field of general-purpose AI will finally begin to mature." https://garymarcus.substack.com/p/ai-still-lacks-common-sense-70-years (fetched). FLAG: every 2024-or-later item on his own writing list (https://cs.nyu.edu/~davise/generalWriting.html, fetched) is co-written with Marcus; I found no solo item. A 2019 AI Impacts conversation (https://aiimpacts.org/conversation-with-ernie-davis/, seen in search result) is solo but too early. Quota tags: skeptic: YES. "not … any time in the foreseeable future" (2014-10-23, published 2015-03, https://cs.nyu.edu/~ed1/papers/Bostrom.pdf). | non_us_uk: no (NYU, United States) | philosopher: no (computer scientist) | frontier_lab: no Verdict: passes C1 (primary, clear). C2 passes only if co-written statements count; flag for John. If a solo statement is required, C2 fails. Queries: (1) Ernest Davis "Ethical guidelines for a superintelligence" Artificial Intelligence 2015 Bostrom; (2) "Ernest Davis" NYU AGI OR "human-level" OR superintelligence 2024 OR 2025 essay OR interview. (2 of 4 used.)

Toby Walsh

Tier checked: core (C1, C2 and quota tags only) C1 earliest: 2016-02-20 (arXiv 1602.06462, "The Singularity May Never Be Near") | timing | "Like many of my colleagues working in AI, I predict we are just 30 or 40 years away from this event." | https://ar5iv.labs.arxiv.org/html/1602.06462 (fetched) | primary. Confirmed. In context, "this event" means machines matching or exceeding human intelligence; he rejects the runaway "singularity" while expecting "machines that exhibit super-human levels of intelligence". C2 latest: 2025-07-04 (China Daily Hong Kong column, "Will artificial intelligence outsmart humankind?") | it is "scientifically plausible" that machines will match and surpass human intelligence within our children's lifetimes, possibly our own (fetched summary; only "scientifically plausible" came back as his verbatim words) | https://www.chinadailyhk.com/article/615214 (fetched) | primary. Confirmed. Another verbatim line from it: "it's not intelligence that's the risk. It's power". Quota tags: non_us_uk: YES. University of New South Wales and Data61 (formerly NICTA), Sydney, Australia, as given in the 2016 paper (2016-02-20, https://ar5iv.labs.arxiv.org/html/1602.06462). | skeptic: NEAR MISS, does not apply. He doubts the singularity, but his pre-2017 forecast is human-level AI in "30 or 40 years" (about 2046 to 2056), which is within decades. That fails the test, as a "2029" would. | philosopher: no | frontier_lab: no Verdict: passes C1 and C2 (both primary, both confirmed). Tag: non_us_uk only. Queries: none (all done with fetches). (0 of 4 used.)

Summary

Person C1 C2 C3 Tags
Shanahan pass (2014, secondhand) not found 0 to 1 found frontier_lab (DeepMind 2017 to present); philosopher near miss
Goertzel pass (2006) pass (2024-03) none found non_us_uk partial (Hong Kong about 2011 on, unconfirmed)
Wooldridge marginal (2017-10, "conscious machines") pass (2025-11) none found skeptic probable, but only 2017-10 evidence
Pearl thin (2011, possible paraphrase) pass (2024-09) 0 to 1 found none (philosopher near miss)
Tegmark pass (2014, co-written) pass (2025-06) 3 (one is his own op-ed) none
Mitchell not found from 2017 or earlier pass (2025-09) (other helper) philosopher likely, needs a source
Etzioni pass (2014-12) not found (other helper) skeptic
Davis pass (2014-10, primary) pass only if co-written counts (2025-07) (other helper) skeptic
Walsh pass (2016-02) pass (2025-07) (other helper) non_us_uk

Systematic caveat: the UK national outlets (BBC, Guardian, Independent, Telegraph) and New Scientist and Wired are blocked here. The C3 gaps for Shanahan and Wooldridge, both UK-based, are probably an artefact of that and need a manual archive check before either is dropped.

Rendered from docs/panel-evidence-2026-10/a13-core-new-d2.md at version a7bd17a.