Urges abandoning the unbounded AGI “everything-machine” goal and building well-defined, well-scoped AI instead; this does not establish a blanket ban on all AI research or all current frontier models. [1][2][3][4][5][6][7][8][9]
Initial statement. Private scoped research synthesis; no public record, portrait crop or human approval exists.
Source [1] · Coauthored research article
The TESCREAL bundle: Eugenics and the promise of utopia through artificial general intelligence
Attributed to: Timnit Gebru · First Monday
Published
“…urge researchers and practitioners to abandon this goal in lieu of building well-defined, well-scoped systems that prioritize people’s safety.”
Section 8 Conclusion, final paragraph, clause beginning “urge researchers” · View original source ↗
Source context. Coauthored conclusion, section 8 final paragraph. The antecedent is AGI as an unbounded everything-machine/god, not every AI model. Section 7 rejects unscoped systems as inherently untestable, and positively recommends well-defined narrow systems. Article online publication is14April2024 (journal issue is dated1April2024); received 21 September 2023 and accepted 10 March 2024 are editorial dates, not inferred statement dates.
Scope
Development of unbounded AGI described as an everything-machine; well-defined, well-scoped systems are the recommended alternative.
Conditions
Abandon the unbounded AGI goal in favor of well-scoped systems; no eventual resumption of that goal is proposed.
Stated rationale
The authors argue that an unscoped system cannot be adequately tested for safety.
Source details
primary · Independently inspected original full article/PDF through web tool; Grok is not verification. Publication date used rather than inventing composition or recording date. Journal article landing page https://firstmonday.org/ojs/index.php/fm/article/view/13636 explicitly says Published2024-04-14. The issue and full-text footer say Volume29 Number4 -1April2024, an issue date. Use article online publicationApr14; preserve discrepancy; no separate composition or statement date invented.
Discovered 6 Oct 2026.
Published 14 Apr 2024.
Accessed 6 Oct 2026.
Source [2] · Coauthored research article
The TESCREAL bundle: Eugenics and the promise of utopia through artificial general intelligence
Attributed to: Timnit Gebru · First Monday
Published
“…urge researchers and practitioners to abandon this goal in lieu of building well-defined, well-scoped systems that prioritize people’s safety.”
Section 8 Conclusion, final paragraph, clause beginning “urge researchers” · View original source ↗
Source context. Coauthored conclusion, section 8 final paragraph. The antecedent is AGI as an unbounded everything-machine/god, not every AI model. Section 7 rejects unscoped systems as inherently untestable, and positively recommends well-defined narrow systems. Article online publication is14April2024 (journal issue is dated1April2024); received 21 September 2023 and accepted 10 March 2024 are editorial dates, not inferred statement dates.
Scope
Development of well-defined, well-scoped AI systems as the alternative to unbounded AGI.
Conditions
Prioritize people’s safety; no explicit rate target or development halt for these bounded systems.
Stated rationale
Well-scoped systems permit appropriate safety tests.
Source details
primary · Independently inspected original full article/PDF through web tool; Grok is not verification. Publication date used rather than inventing composition or recording date. Journal article landing page https://firstmonday.org/ojs/index.php/fm/article/view/13636 explicitly says Published2024-04-14. The issue and full-text footer say Volume29 Number4 -1April2024, an issue date. Use article online publicationApr14; preserve discrepancy; no separate composition or statement date invented.
Discovered 6 Oct 2026.
Published 14 Apr 2024.
Accessed 6 Oct 2026.
Source [3] · Edited original interview transcript
Timnit Gebru Believes There Is No ‘Existential Threat’ From AI
Attributed to: Timnit Gebru · WIRED / Lauren Goode
Published
“I have to be actively making space for the kinds of models that I think should be built and building them”
Answer to “What would you say is the biggest part of your own thinking ... that has evolved”, transcript paragraph beginning “I have to” · View original source ↗
Source context. Original interview publication September 29, 2026; actual recording date not established. Interview criticizes AGI hype and emphasizes alternative models and human agency. Does not explicitly restate an enduring AGI prohibition or propose AGI reopening. Same interview also endorses data/labor protections and describes their predicted slowdown effect before deployment. Regulation alone does not establish a development pace directive; no explicit reopening of the unbounded AGI goal appears.
Scope
Alternative AI models Gebru believes should be built; specific model/capability coverage is unstated.
Conditions
No pace or AGI reopening policy is stated in this passage.
Stated rationale
Make space for constructive work rather than react to industry narratives.
Source details
primary · Independently inspected original full article/PDF through web tool; Grok is not verification. Publication date used rather than inventing composition or recording date.
Discovered 6 Oct 2026.
Published 29 Sep 2026.
Accessed 6 Oct 2026.
Source [4] · Original event report with attributed quotations
Timnit Gebru Says Artificial Intelligence Needs to Slow Down
Attributed to: Timnit Gebru · WIRED / Max G. Levy
Published
“The incentive structure is not such that you slow down, first of all, think about how you should approach research”
Paragraph beginning “The incentive structure”, following Stochastic Parrots account · View original source ↗
Source context. Publisher’s own report of November9REWIRED conversation; exact quotation attributed toGebru. Fullspokenrecording not independently recovered. Source also includes safety-before-deployment instruction, distinct from developmenthalt.
Scope
AI research incentives and the race toward larger language models in2021.
Conditions
Slow and deliberate to examine when and how AI should be built; not a fixed moratorium or enduring prohibition.
Stated rationale
Avoid harms and make accountable choices.
Source details
secondary · Original publisher independently opened; source context, attribution and publication date inspected. Rush transcript caveat retained for DemocracyNow.
Discovered 6 Oct 2026.
Published 9 Nov 2021.
Accessed 6 Oct 2026.
Source [5] · Original interview rush transcript
Deep Unlearning: Timnit Gebru on AI Hype, Ethics & Algorithmic Racial Bias
Attributed to: Timnit Gebru · Democracy Now!
Published
“The path we’re on right now was never a preordained path”
Gebru’s response concerning DAIR’s data-workers project, paragraph beginning “And so, for me” · View original source ↗
Source context. Retrospective; no development restart directive.
Scope
Alternative technology paths and data-worker conditions.
Conditions
No development duration or reopening rule stated.
Stated rationale
Avoid harms and make accountable choices.
Source details
primary · Original publisher independently opened; source context, attribution and publication date inspected. Rush transcript caveat retained for DemocracyNow.
Discovered 6 Oct 2026.
Published 13 Aug 2026.
Accessed 6 Oct 2026.
Source [6] · Original interview rush transcript
The Real Existential Risk Is AI CEOs Avoiding Oversight
Attributed to: Timnit Gebru · Democracy Now!
Published
“Do not allow companies to put out unreliable products into the world”
Final policy answer before eugenics question, paragraph beginning “So, for me, it’s very clear” · View original source ↗
Source context. Deployment instruction; not capability-development reopening. Earlier in the same October1originalinterview, AmyGoodman explicitly asks about BernieSanders’s effort to ban artificialsuperintelligence, describing Gebru’s recent article as calling his bill well-meaning but misguided. Gebru answers thatSanders is leading the left astray and criticizes affiliations, regulatorycapture, and promoting a superintelligence framing in place of holding corporations accountable for faulty products. This is material contrary context to a blanket interpretation ofSTOP. It does not expressly endorse continued unboundedAGIdevelopment or adopt a safety-conditioned reopeningpolicy.
Scope
Deployment of unreliable products.
Conditions
Tests and guardrails required.
Stated rationale
Avoid harms and make accountable choices.
Source details
primary · Original publisher independently opened; source context, attribution and publication date inspected. Rush transcript caveat retained for DemocracyNow.
Discovered 6 Oct 2026.
Published 1 Oct 2026.
Accessed 6 Oct 2026.
Source [7] · Coauthored response statement
Statement from the listed authors of Stochastic Parrots on the “AI pause” letter
Attributed to: Timnit Gebru · Distributed AI Research Institute
Published
“We should be building machines that work for us”
Paragraph after accountability discussion, beginning “We should be building machines” · View original source ↗
Source context. Root agent independently inspected complete original page in existing authenticated Brave. Visible bylineTimnitGebru,EmilyBender,AngelinaMcMillanMajor; visible dateMarch30,2023 (GrokMarch31lead corrected). Responds toFLIsixmonth letter; rejects utopia/apocalypse framing, emphasizes ongoing harms and regulation protecting people. The quoted sentence contrasts useful machines with adapting society to be machine-readable, followed by rejecting the race as preordained.
Scope
Alternative AI research direction and accountability for existing corporate practices.
Conditions
No adopted development suspension duration or AGI reopening policy stated.
Stated rationale
Present harms and corporate accountability.
Source details
primary · Primary original independently inspected by codex-root-20261006 through existingBrave; directfetch was403. OriginaldateMarch30overridesunverifiedGrokMarch31. Noauthorcompositiondate invented.
Discovered 6 Oct 2026.
Published 30 Mar 2023.
Accessed 6 Oct 2026.
Source [8] · Original social thread
Critical thread on Peter Thiel’s2013EffectiveAltruismkeynote
Attributed to: Timnit Gebru · Timnit Gebru / X
Published
“…even though this whole AGI thing is THE MOST important thing to work on potentially.”
Post1996768491601645906 in thread beginning1996756697327849719; preceding1996768478683082766 and following1996768652889411806 · View original source ↗
Source context. Root independently inspected complete originalthread in existingBrave. Thread begins withGebru’s displeasure watchingThiel2013EAkeynote; precedingpost describesThiel’s opposition to measurement for immeasurable problems; targetcontinues reported/sarcastic account; followingpost criticizesEAquantification. DOMdatetime2025-12-05T02:28:01Z corresponds to visible6:28PMDec4Pacific. AGIpriority language is not her endorsed recommendation.
Scope
Critical account of PeterThiel’s2013EffectiveAltruismkeynote.
Conditions
Sarcastic/reported account ofThiel, notGebru’s own development instruction.
Stated rationale
Critical discussion ofEApriorities and quantification.
Source details
primary · Completeoriginalthreadreadbycodex-root-20261006throughauthenticatedexistingBrave; exactDOMdatetime and localdisplaydate checked. Initialweblookupblocked, preservedprivately; noCONTINUEinferred.
Discovered 6 Oct 2026.
Published 5 Dec 2025.
Accessed 6 Oct 2026.
Source [9] · Original interview rush transcript
The Real Existential Risk Is AI CEOs Avoiding Oversight
Attributed to: Timnit Gebru · Democracy Now!
Published
“I think that Bernie Sanders is really leading the left astray”
Response immediately following AmyGoodman’s question “can you talk about Senator Bernie Sanders’ effort to ban artificial superintelligence?” · View original source ↗
Source context. Deployment instruction; not capability-development reopening. Earlier in the same October1originalinterview, AmyGoodman explicitly asks about BernieSanders’s effort to ban artificialsuperintelligence, describing Gebru’s recent article as calling his bill well-meaning but misguided. Gebru answers thatSanders is leading the left astray and criticizes affiliations, regulatorycapture, and promoting a superintelligence framing in place of holding corporations accountable for faulty products. This is material contrary context to a blanket interpretation ofSTOP. It does not expressly endorse continued unboundedAGIdevelopment or adopt a safety-conditioned reopeningpolicy.
Scope
Criticism of Sanders’s ASI-ban framing and political direction.
Conditions
No endorsed development go-ahead or AGI reopening policy stated.
Stated rationale
Opposes industry marketing/regulatorycapture through fictional superintelligence framing.
Source details
primary · Original publisher independently opened; source context, attribution and publication date inspected. Rush transcript caveat retained for DemocracyNow.
Discovered 6 Oct 2026.
Published 1 Oct 2026.
Accessed 6 Oct 2026.
Editorial assessment. Researcher proposal, awaiting separate independent adversarial review and exact human editorial approval. No new public identity or stance has been integrated. Earlier in the same October1originalinterview, AmyGoodman explicitly asks about BernieSanders’s effort to ban artificialsuperintelligence, describing Gebru’s recent article as calling his bill well-meaning but misguided. Gebru answers thatSanders is leading the left astray and criticizes affiliations, regulatorycapture, and promoting a superintelligence framing in place of holding corporations accountable for faulty products. This is material contrary context to a blanket interpretation ofSTOP. It does not expressly endorse continued unboundedAGIdevelopment or adopt a safety-conditioned reopeningpolicy.