MIT Technology Review · Sep 22, 2026
Timnit Gebru and Emily M. Bender argue in MIT Technology Review that AI hype misdirects
Markdown · JSON · Original source
Timnit Gebru and Emily M. Bender’s essay was published on September 22, 2026. They argue that companies’ claims about AI breakthroughs and dangerous “” receive breathless coverage, while later scrutiny by experts gets less attention. In their account, the hacking incidents point to ’s negligence and failures to follow basic security practices, not models acting on their own; mathematicians also questioned claims about mathematical breakthroughs. The authors argue that math and programming are attractive showcases partly because their answers can be checked, making system outputs easier to evaluate without paying workers to annotate them. They warn that describing products as “rogue models” or “superintelligence” shifts attention from company actions and accountability. They also argue that this framing diverts attention from harms they name, including data centers’ effects on climate, nearby residents’ asthma, electricity bills, and water use. Their conclusion is that policymakers and communities should take to hear independent experts and contextualize corporate claims, rather than make decisions under pressure. The essay does not establish how much this framing changes policy or public decisions.
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
MIT Technology Review · Article
Source details
Original source date: Sep 22, 2026 (2026-09-22T11:04:51Z).
Date basis: Date from the publisher’s own records, through its API.
Added to Escape Meridian: 2026-11-09T00:23:31.887Z. First observed: 2026-11-09T00:07:41.215Z. Retrieved: 2026-11-09T00:09:34.522Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
Text as retrieved on November 9, 2026; originally published September 22, 2026. The essay presents the authors’ arguments and reports others’ claims; the supplied text does not independently establish the disputed company claims or the extent of their effects on policy and public decisions.
Open question
What evidence would show whether coverage of corporate AI claims is diverting policymakers’ attention from company accountability and data-center harms?
Source-stated
Gebru and Bender argue that intense coverage of company AI claims often precedes expert scrutiny, whose differing accounts receive less attention.
Once there is time for experts in the relevant fields to examine what happened, a very different story emerges, but one that gets less media attention.
Source-stated
The authors say cybersecurity experts characterized the hacking incidents as failures of basic security practices, rather than models acting independently.
cybersecurity experts say the story is more about OpenAI’s negligence and failure to adopt basic, established security practices than about “models gone rogue”
Source-stated
Gebru and Bender argue that math and programming are appealing fields for evaluating system outputs because answers can be verified, reducing the need for paid annotation.
The latter makes math and coding problems easier to tune systems for, since system output (sequences of likely words or pieces of computer code) can be evaluated without having to pay data workers to look at and annotate each one.
Source-stated
The authors argue that framing products as “superintelligence” or “rogue models” assigns agency to products and helps companies evade accountability.
Describing them as “superintelligence” or “rogue models” ascribes agency to products rather than to the companies building them.
Source-stated
Gebru and Bender conclude that policymakers and communities should take time to consult independent experts and contextualize corporate claims.
Wise decision-making, by policymakers and communities, demands time to hear from independent experts and contextualize corporate claims.
Source and contributor context
MIT Technology Review · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Superintelligence
A hypothetical AI far more capable than people across most intellectual tasks.
OpenAI
OpenAI develops AI models and products and publishes research, release announcements and evaluations.
TIME
TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.
Editorial provenance
Version 1. Publication ID: c242142e-d21e-1386-448c-a1e903ce5ff2.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-09T00:19:10.388Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model.
Observation: 935fc53063ab9dbb1b9e57febf7d2420e45ed5e70a7118171cac8510ee1b30ed. Source text hash: cad2b5997dbf032459555d8148200cda6a8b4c91a4dbf8f8f12df1160443260b.
Revision and source-check history
Version 1 · 2026-11-09T00:23:31.887Z
Timnit Gebru and Emily M. Bender argue in MIT Technology Review that AI hype misdirects
Timnit Gebru and Emily M. Bender’s essay was published on September 22, 2026. They argue that companies’ claims about AI breakthroughs and dangerous “” receive breathless coverage, while later scrutiny by experts gets less attention. In their account, the hacking incidents point to ’s negligence and failures to follow basic security practices, not models acting on their own; mathematicians also questioned claims about mathematical breakthroughs. The authors argue that math and programming are attractive showcases partly because their answers can be checked, making system outputs easier to evaluate without paying workers to annotate them. They warn that describing products as “rogue models” or “superintelligence” shifts attention from company actions and accountability. They also argue that this framing diverts attention from harms they name, including data centers’ effects on climate, nearby residents’ asthma, electricity bills, and water use. Their conclusion is that policymakers and communities should take to hear independent experts and contextualize corporate claims, rather than make decisions under pressure. The essay does not establish how much this framing changes policy or public decisions.
Original source
MIT Technology Review · Article
Source details
Original source date: Sep 22, 2026 (2026-09-22T11:04:51Z).
Date basis: Date from the publisher’s own records, through its API.
Added to Escape Meridian: 2026-11-09T00:23:31.887Z. First observed: 2026-11-09T00:07:41.215Z. Retrieved: 2026-11-09T00:09:34.522Z.
Limits
Text as retrieved on November 9, 2026; originally published September 22, 2026. The essay presents the authors’ arguments and reports others’ claims; the supplied text does not independently establish the disputed company claims or the extent of their effects on policy and public decisions.
Open question
What evidence would show whether coverage of corporate AI claims is diverting policymakers’ attention from company accountability and data-center harms?
Source-stated
Gebru and Bender argue that intense coverage of company AI claims often precedes expert scrutiny, whose differing accounts receive less attention.
Once there is time for experts in the relevant fields to examine what happened, a very different story emerges, but one that gets less media attention.
Source-stated
The authors say cybersecurity experts characterized the hacking incidents as failures of basic security practices, rather than models acting independently.
cybersecurity experts say the story is more about OpenAI’s negligence and failure to adopt basic, established security practices than about “models gone rogue”
Source-stated
Gebru and Bender argue that math and programming are appealing fields for evaluating system outputs because answers can be verified, reducing the need for paid annotation.
The latter makes math and coding problems easier to tune systems for, since system output (sequences of likely words or pieces of computer code) can be evaluated without having to pay data workers to look at and annotate each one.
Source-stated
The authors argue that framing products as “superintelligence” or “rogue models” assigns agency to products and helps companies evade accountability.
Describing them as “superintelligence” or “rogue models” ascribes agency to products rather than to the companies building them.
Source-stated
Gebru and Bender conclude that policymakers and communities should take time to consult independent experts and contextualize corporate claims.
Wise decision-making, by policymakers and communities, demands time to hear from independent experts and contextualize corporate claims.
Source and contributor context
MIT Technology Review · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Superintelligence
A hypothetical AI far more capable than people across most intellectual tasks.
OpenAI
OpenAI develops AI models and products and publishes research, release announcements and evaluations.
TIME
TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.
Editorial provenance
Version 1. Publication ID: c242142e-d21e-1386-448c-a1e903ce5ff2.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-09T00:19:10.388Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model.
Observation: 935fc53063ab9dbb1b9e57febf7d2420e45ed5e70a7118171cac8510ee1b30ed. Source text hash: cad2b5997dbf032459555d8148200cda6a8b4c91a4dbf8f8f12df1160443260b.
uncertain · 2026-11-09T00:19:10.388Z · anthropic/claude-opus-5.5
Publication: c242142e-d21e-1386-448c-a1e903ce5ff2
attribution — pass: The reading attributes the essay to Timnit Gebru and Emily M. Bender in MIT Technology Review, which matches the author bio in the text. It keeps company and expert claims attributed, for example the cybersecurity experts' view. The page is one argued essay, not a listing.
Timnit Gebru is executive director of DAIR and author of the forthcoming book Deep Unlearning
summary — uncertain: The summary refers to "the hacking incidents" without saying what they were, such as the OpenAI–Hugging Face incident, so a newcomer lacks essential context. It also softens the authors' force. It omits their charge that superintelligence claims rest on ideologies of transhumanism, eugenics and wishful thinking. It omits mathematicians' plagiarism and misconduct accusations and the critique of Sanders's bill as a misdirected response. Smallest repair: name the OpenAI–Hugging Face incident and add one sentence on the ideology charge.
Rather, they are narratives based in ideologies of transhumanism, eugenics, and wishful thinking about imagined future digital humans.
date — pass: September 22, 2026 matches the UTC day of originalPublished, 2026-09-22T11:04:51Z. The reading does not use the later sidebar items or present any forecast as an outcome. The evidence does not mark the page as edited after publication, although modifiedAt shows a change on Sept 25. No reading claim depends on text that is plainly from a later edit.
support — pass: Each claim matches its quoted excerpt and stays attributed to the authors or to the experts they cite. The question is open and non-leading. The limitation correctly notes that the essay does not establish the effect on policy.
Describing them as “superintelligence” or “rogue models” ascribes agency to products rather than to the companies building them.
media — pass: No media was selected and no verified candidates were supplied, which is appropriate.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
A hypothetical AI far more capable than people across most intellectual tasks.
Definition referenceOpenAI develops AI models and products and publishes research, release announcements and evaluations.
Explore OpenAITIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.
Explore TIMEA hypothetical AI far more capable than people across most intellectual tasks.
Definition referenceOpenAI develops AI models and products and publishes research, release announcements and evaluations.
Explore OpenAITIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.
Explore TIME