# Karen Hao traces OpenAI’s shift in MIT Technology Review [Canonical reading](https://escapemeridian.com/readings/karen-hao-traces-openai-s-shift-in-mit-technology-review-30c9999c) On May 19, 2025, Karen Hao recounts in MIT Technology Review how reporting on OpenAI began with a three-day visit in 2019 and grew into a profile of a lab changing its structure, funding and approach to openness. Greg Brockman and Ilya Sutskever argued that building artificial general intelligence (AGI)—which OpenAI defined as software matching or exceeding human performance on most economically valuable tasks—could help solve complex problems and distribute economic benefits. Hao’s account contrasts those promises with their difficulty giving concrete examples of how AGI would help, address environmental costs or share its gains. Brockman said falling behind could undermine OpenAI’s mission; Hao argues that the need to lead put pressure on research and encouraged heavy use of compute and data, with environmental and consent costs. Her profile concluded that OpenAI’s public commitments to transparency and collaboration diverged from its operations. An email from Sam Altman, as Hao recounts it, proposed responding through public messaging rather than changes to internal practices. The excerpt leaves open how OpenAI would distribute benefits or make its stated commitments concrete. 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 Original source date: May 19, 2025 (2025-05-19T09:00:00Z). Date basis: Date from the publisher’s own records, through its API. Added to Escape Meridian: 2026-11-14T03:48:28.662Z. First observed: 2026-10-30T08:22:22.359Z. Retrieved: 2026-10-30T08:23:46.164Z. - [Inside the story that enraged OpenAI](https://www.technologyreview.com/2025/05/19/1116614/hao-empire-ai-openai/) ## Limits Text as retrieved on October 30, 2026; originally published May 19, 2025. The excerpt recounts Hao’s reporting and interviews, but does not establish whether OpenAI’s proposed distribution of AGI’s benefits or its stated commitments were fulfilled. The basis for measuring those outcomes is not provided. ## Open question What concrete approach, if any, would OpenAI use to distribute AGI’s benefits, beyond the possibilities Brockman floated? ## Source-stated Hao recounts that Brockman and Sutskever presented AGI as a way to help solve complex problems and distribute its economic benefits. > Our mission is to ensure that AGI benefits all of humanity. And the way we want to do that is: Build AGI and distribute its economic benefits. - [Supporting source](https://www.technologyreview.com/2025/05/19/1116614/hao-empire-ai-openai/) ## Source-stated Hao argues that the pressure to lead shaped OpenAI’s conduct and drove resource use, including compute and data, with environmental and consent costs. > It justified OpenAI’s consumption of an unfathomable amount of resources: both compute, regardless of its impact on the environment; and data, the amassing of which couldn’t be slowed by getting consent or abiding by regulations. - [Supporting source](https://www.technologyreview.com/2025/05/19/1116614/hao-empire-ai-openai/) ## Source-stated Hao’s 2020 profile concluded that OpenAI’s public commitments and its operations diverged. > There is a misalignment between what the company publicly espouses and how it operates behind closed doors - [Supporting source](https://www.technologyreview.com/2025/05/19/1116614/hao-empire-ai-openai/) ## Source-stated Hao recounts that Altman proposed addressing criticism through public messaging rather than changes to internal practices. > This could be smoothed over not with changes to its internal practices but some tuning of OpenAI’s public messaging. - [Supporting source](https://www.technologyreview.com/2025/05/19/1116614/hao-empire-ai-openai/) ## 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. ## Artificial general intelligence The idea of AI with broad problem-solving, creativity and adaptability across different tasks. - [Definition reference](https://developers.google.com/machine-learning/glossary#artificial-general-intelligence) ## OpenAI OpenAI develops AI models and products and publishes research, release announcements and evaluations. - [Explore this context](/entities/openai) - [Reference](https://openai.com/news/) ## Greg Brockman Greg Brockman publishes personal essays about engineering, AI research and organizational work. - [Explore this context](/entities/greg-brockman) - [Reference](https://blog.gregbrockman.com/) ## Sam Altman Sam Altman publishes essays about AI and technology on his personal blog, separate from OpenAI announcements. - [Explore this context](/entities/sam-altman) - [Reference](https://blog.samaltman.com/) - [Reference](https://openai.com/our-structure/) - [Reference](https://blog.samaltman.com/posts.atom) ## Editorial provenance Version 1. Publication ID: 3b7828c1-c6e4-e90f-defc-ae8412a17818. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-14T03:43:24.829Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: bridge\_admitted\_by\_model. Observation: 88ee37b7099d0e998c1b872acc09a03a2b3b7c1fb0f5d0b52c583438e18e1c2e. Source text hash: ed0e2097f59c4ad0b335d04aa4eb60915f5daac569f5027ba325fff9d712f18c. ## Revision and source-check history ### Version 1 — 2026-11-14T03:48:28.662Z On May 19, 2025, Karen Hao recounts in MIT Technology Review how reporting on OpenAI began with a three-day visit in 2019 and grew into a profile of a lab changing its structure, funding and approach to openness. Greg Brockman and Ilya Sutskever argued that building artificial general intelligence (AGI)—which OpenAI defined as software matching or exceeding human performance on most economically valuable tasks—could help solve complex problems and distribute economic benefits. Hao’s account contrasts those promises with their difficulty giving concrete examples of how AGI would help, address environmental costs or share its gains. Brockman said falling behind could undermine OpenAI’s mission; Hao argues that the need to lead put pressure on research and encouraged heavy use of compute and data, with environmental and consent costs. Her profile concluded that OpenAI’s public commitments to transparency and collaboration diverged from its operations. An email from Sam Altman, as Hao recounts it, proposed responding through public messaging rather than changes to internal practices. The excerpt leaves open how OpenAI would distribute benefits or make its stated commitments concrete. ## Original source MIT Technology Review · Article Original source date: May 19, 2025 (2025-05-19T09:00:00Z). Date basis: Date from the publisher’s own records, through its API. Added to Escape Meridian: 2026-11-14T03:48:28.662Z. First observed: 2026-10-30T08:22:22.359Z. Retrieved: 2026-10-30T08:23:46.164Z. - [Inside the story that enraged OpenAI](https://www.technologyreview.com/2025/05/19/1116614/hao-empire-ai-openai/) ## Limits Text as retrieved on October 30, 2026; originally published May 19, 2025. The excerpt recounts Hao’s reporting and interviews, but does not establish whether OpenAI’s proposed distribution of AGI’s benefits or its stated commitments were fulfilled. The basis for measuring those outcomes is not provided. ## Open question What concrete approach, if any, would OpenAI use to distribute AGI’s benefits, beyond the possibilities Brockman floated? ## Source-stated Hao recounts that Brockman and Sutskever presented AGI as a way to help solve complex problems and distribute its economic benefits. > Our mission is to ensure that AGI benefits all of humanity. And the way we want to do that is: Build AGI and distribute its economic benefits. - [Supporting source](https://www.technologyreview.com/2025/05/19/1116614/hao-empire-ai-openai/) ## Source-stated Hao argues that the pressure to lead shaped OpenAI’s conduct and drove resource use, including compute and data, with environmental and consent costs. > It justified OpenAI’s consumption of an unfathomable amount of resources: both compute, regardless of its impact on the environment; and data, the amassing of which couldn’t be slowed by getting consent or abiding by regulations. - [Supporting source](https://www.technologyreview.com/2025/05/19/1116614/hao-empire-ai-openai/) ## Source-stated Hao’s 2020 profile concluded that OpenAI’s public commitments and its operations diverged. > There is a misalignment between what the company publicly espouses and how it operates behind closed doors - [Supporting source](https://www.technologyreview.com/2025/05/19/1116614/hao-empire-ai-openai/) ## Source-stated Hao recounts that Altman proposed addressing criticism through public messaging rather than changes to internal practices. > This could be smoothed over not with changes to its internal practices but some tuning of OpenAI’s public messaging. - [Supporting source](https://www.technologyreview.com/2025/05/19/1116614/hao-empire-ai-openai/) ## 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. ## Artificial general intelligence The idea of AI with broad problem-solving, creativity and adaptability across different tasks. - [Definition reference](https://developers.google.com/machine-learning/glossary#artificial-general-intelligence) ## OpenAI OpenAI develops AI models and products and publishes research, release announcements and evaluations. - [Explore this context](/entities/openai) - [Reference](https://openai.com/news/) ## Greg Brockman Greg Brockman publishes personal essays about engineering, AI research and organizational work. - [Explore this context](/entities/greg-brockman) - [Reference](https://blog.gregbrockman.com/) ## Sam Altman Sam Altman publishes essays about AI and technology on his personal blog, separate from OpenAI announcements. - [Explore this context](/entities/sam-altman) - [Reference](https://blog.samaltman.com/) - [Reference](https://openai.com/our-structure/) - [Reference](https://blog.samaltman.com/posts.atom) ## Editorial provenance Version 1. Publication ID: 3b7828c1-c6e4-e90f-defc-ae8412a17818. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-14T03:43:24.829Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: bridge\_admitted\_by\_model. Observation: 88ee37b7099d0e998c1b872acc09a03a2b3b7c1fb0f5d0b52c583438e18e1c2e. Source text hash: ed0e2097f59c4ad0b335d04aa4eb60915f5daac569f5027ba325fff9d712f18c. ### uncertain — 2026-11-14T03:43:24.829Z — anthropic/claude-opus-5.5 Publication: 3b7828c1-c6e4-e90f-defc-ae8412a17818 attribution — pass: The reading attributes the account to Karen Hao in MIT Technology Review and attributes statements to Brockman, Sutskever and Altman as Hao recounts them. This is one piece, a book excerpt with an editor's introduction. The reading could optionally name Empire of AI, but the attribution is not wrong. > From the book Empire of AI: Dreams and Nightmares in Sam Altman’s OpenAI, by Karen Hao summary — pass: The reading clearly identifies the actors and defines AGI using OpenAI's definition. It conveys Hao's central argument that the need to lead drove OpenAI's conduct, and it states her profile's conclusion and Altman's messaging response. A lay reader can follow it. It leaves out that OpenAI stopped speaking to Hao for three years, but that omission does not distort the meaning. date — pass: May 19, 2025 matches the UTC day of originalPublished. The 2019 visit and 2020 profile dates match the source. The sidebar items on the retrieved page, such as the AlphaGo ten-years piece, appear to be later additions. The reading neither uses nor relies on them. The reading uses no hindsight framing. The page is a publisher page, not an archive copy. support — uncertain: The four claims match their excerpts. However, the summary says Brockman and Sutskever had difficulty giving concrete examples of how AGI would help. In the source, Brockman did offer examples: climate change, medicine, and a friend's rare disorder. What Hao says she failed to get was concrete detail on what they were building, and Sutskever gave no examples of AGI offsetting environmental costs. The smallest repair is to say they struggled to give concrete details of what they were building, or of how AGI would offset environmental costs. > He offered two examples that had become dogma among AGI believers. Climate change. media — pass: No media was selected, and no verified candidates were supplied, so this is appropriate. ## Recorded references Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter. No public references recorded here yet. [Reference data (JSON)](/readings/30c9999ca199ea209c27c32172dcfaffc7c6a55aed20d9e51346163e66c75c31/references.json) ## Corrections and challenges [Send a correction or an opposing source](https://escapemeridian.com/corrections/) [JSON record](https://escapemeridian.com/readings/karen-hao-traces-openai-s-shift-in-mit-technology-review-30c9999c.json)