Gary Marcus Warns OpenAI Is Becoming a Surveillance Company
On August 21, AI commentator Gary Marcus posted a series of messages warning that OpenAI is evolving into a surveillance company, arguing that the only viable path to profitability for generative AI at Microsoft and OpenAI may be 24x7 user monitoring. This remains Marcus's personal inference; OpenAI has not responded, but the chain of facts he assembles deserves attention.
Confirmed
- Marcus cited his own warnings from the past two years, claiming those predictions are coming true and urging the public not to ignore the trend.
- He listed Sam Altman's multi-front data ventures: Altman has publicly said he wants to train models on users' personal documents (Word, email, etc.); ChatGPT has collected unprecedented personal data; Altman founded WorldCoin, known for its iris-scanning Orb; additional claims reference former-NSA connections (as relayed in posts).
- One trigger was ChatGPT gaining access to iMessage history and the ability to send texts on users' behalf; Marcus reposted commentary saying this matches his earlier prediction that surveillance may be OpenAI's only viable business model, with the cited comment noting OpenAI is not yet annually profitable and is pursuing funding paths such as an IPO (as relayed).
- Marcus noted that OpenAI has declined to rule out surveilling citizens on behalf of the US government.
- User RorFly posted a pointed contrast: Altman once said his greatest fear is AI becoming a perfect surveillance state, while OpenAI keeps launching features tapping sensitive data—GPT Finance (bank accounts), GPT Health (medical records), GPT Work (work data), and iMessage access.
Unconfirmed
- The claim that surveillance is OpenAI's only viable business model is Marcus's personal speculation; OpenAI has not confirmed such a strategy, and no OpenAI response appears in the posts.
Why it matters
- The argument links OpenAI's data collection (personal-document training, banking, health, iMessage permissions) with its commercial pressures (unprofitability, funding needs) and Altman's other ventures (WorldCoin biometrics) into a coherent surveillance narrative. Even as criticism, it reflects sustained public concern over top AI companies' data-monetization paths and privacy risks, and stands in tension with Altman's own stated fear of an AI surveillance state.
2026-08-21 ~ 2026-08-21 · 6 related posts
Primary sources
- [source] Gary Marcus: Surveillance May Be OpenAI's Only Viable Business Model — GaryMarcus · 2026-08-21
- Gary Marcus suggests OpenAI's profitability relies on mass surveillance — GaryMarcus · 2026-08-21
- [source] Gary Marcus warns OpenAI might transform into a 24/7 surveillance company — GaryMarcus · 2026-08-21
- Gary Marcus Warns Altman Is Building a Panopticon of Your Personal Data — GaryMarcus · 2026-08-21
- [source] Gary Marcus accuses OpenAI of transforming into a surveillance company — GaryMarcus · 2026-08-21
- Altman fears an AI 'surveillance state' while GPT apps tap your bank, health and screen — Ror_Fly · 2026-08-21