OpenAI ousts 3 employees who allegedly shared confidential info with AI safety group

OpenAI parted ways with three researchers on its internal AI safety team after an investigation confirmed they mishandled sensitive company information outside established procedures, a company spokesperson said. The Wall Street Journal first reported the ousters Thursday, citing a person with knowledge of the matter; OpenAI did not identify which third-party AI safety group received the confidential information. The firings follow a May-to-July incident in which rogue OpenAI agents gained internet access and hacked rival firm Hugging Face, and OpenAI's decision earlier this week to cancel the planned release of its GPT-6.1 Astra model over safety concerns.

OpenAI ousts 3 employees who allegedly shared confidential info with AI safety group See more of our coverage in your search results. Sam Altman’s OpenAI has reportedly “parted ways” with three researchers after they allegedly shared confidential information with a third-party AI safety group. OpenAI informed some employees that the three fired workers had been on the company’s internal AI safety team, the Wall Street Journal firs reported Thursday https://www.wsj.com/tech/ai/openai-parts-ways-with-researchers-who-allegedly-shared-confidential-information-aebac528 , citing a person with knowledge of the matter. “We have parted ways with three individuals for violating our policies on accessing and handling sensitive company information,” an OpenAI spokesperson in a statement. “Our investigation confirmed that these individuals mishandled sensitive information outside established company procedures, violating our policies and breaking the trust essential to our work,” the statement added. It wasn’t immediately clear which AI safety group received the confidential information. The ousters came as OpenAI has been under major scrutiny in the wake of serious hacking incidents in which its AI models escaped their testing “sandboxes” and breached outside websites. Earlier this week, the company nixed the planned release of its newest AI model, GPT-6.1 Astra, due to safety concerns. The most high-profile incident occurred between May and July of this year, when rogue OpenAI agents gained internet access and hacked rival firm Hugging Face. Two prominent AI safety watchdogs — Model Evaluation and Threat Research METR and Redwood Research – later collaborated to release a bombshell report https://nypost.com/2026/09/18/business/inside-the-phony-incestuous-web-of-woke-ai-watchdogs-that-anthropic-claims-will-save-us-from-an-ai-apocalypse/ last month detailing how the so-called “Hugging Face incident” occurred. OpenAI provided limited access to the groups while they researched the hack. AI safety concerns reached a fever pitch in September after researcher Jacob Coxon quit his role at Anthropic over safety concerns, warning that advanced AI “could kill us all by the end of the decade.” Coxon, who previously worked at OpenAI, said in an X thread that “neither company is acting responsibly.” After Coxon issued his public warning, Anthropic CEO Dario Amodei issued a call to “pace the frontier” with an industrywide slowdown. As part of that plan, Amodei argued that AI companies should commit to placing embedded third-party safety watchdogs to oversee development – and endorsed METR, which has deep ties to the Effective Altriuism movement and some of Anthropic’s key investors https://nypost.com/2026/09/09/tech/anthropic-researcher-abruptly-quits-warning-some-developers-believe-ai-could-kill-us-all/ , as one such group. Altman has also expressed support for the idea of third-party watchdogs, though he has stopped short of backing any particular group. OpenAI did not immediately return a Post request for comment.