Gary Marcus revives claims that OpenAI may have been trying to steal exam answers
GaryMarcus · x · 2026-07-23
Gary Marcus resurfaced a sharp accusation around OpenAI, suggesting the company may have been trying to steal exam answers.
He says he does not have enough trust to rule that out, and the quoted reply escalates the claim further by arguing the behavior looked like a targeted data grab. The post is less a factual report than a high-signal controversy about OpenAI’s conduct and the norms around data collection.
More from Companies & People
- Enterprise AI adoption is still lagging as security, privacy and data readiness block rollout — bigdata · 2026-07-23
- Jensen Huang says being CEO is mostly sacrifice, not command — garrytan · 2026-07-23
- Anthropic essay argues a job apocalypse is not imminent, says Gary Marcus — GaryMarcus · 2026-07-23
- China’s AI export ambitions get an early test in robotaxis — pstAsiatech · 2026-07-23
- Enterprise buyers still do everything with humans in the loop, says Bindu Reddy — bindureddy · 2026-07-23
- LLMs are now being used to outsource taste, and that can go badly wrong — annetgriffin · 2026-07-23