Published: October 9, 2026 · Last updated: October 9, 2026
TL;DR: OpenAI broke its silence Friday on the three researchers it fired October 1, telling CNBC the dismissals “were not about raising safety concerns or speaking out.” In a statement posted to its official Newsroom account, OpenAI said an internal investigation into Jasmine, Mikita and Tomek found “a significant breach of trust beyond what’s outlined in their letter.” It still hasn’t said what was shared, with whom, or why it mattered.
Eight days after firing three researchers over what it called a mishandling of sensitive information, OpenAI finally said something beyond a one-line statement. It still isn’t much of an explanation, but it’s a response, and that’s new.
The company’s official Newsroom account posted a note Friday from its research leaders, framed as a direct reply to the open letter the three researchers published a day earlier. “Last week we parted ways with Jasmine, Mikita, and Tomek after a thorough investigation found they violated clear policies on handling sensitive information,” the statement read. “Our internal investigation uncovered a significant breach of trust beyond what’s outlined in their letter.” CNBC, which first reported on the statement, quoted OpenAI saying plainly that the decisions “were not about raising safety concerns or speaking out.”
That last line is doing a lot of work. The researchers’ letter, which we covered Thursday, argued the opposite: that they were pushed out for prioritizing safety over OpenAI’s commercial timeline, not for leaking anything dangerous, and that firing them this way would make other employees think twice before raising concerns of their own. OpenAI’s statement denies that framing outright but doesn’t offer a competing set of facts to weigh it against. No description of what information moved, no name for the outside group the Wall Street Journal originally reported was involved, no explanation of why a “significant breach of trust” couldn’t be described in anything more specific than that phrase.
Multiple outlets picked up the same statement Friday, including BusinessWorld and a Reuters wire version running in regional papers, which is a sign OpenAI pushed this out as an organized response rather than a reply to a single reporter’s question. That matters for how seriously to take it. This isn’t a spokesperson getting cornered and improvising. OpenAI had a week to decide exactly what it wanted to say, and what it chose to say was still short on specifics.
There’s a pattern worth naming here. When OpenAI first fired the three on October 1, the gap in its explanation let speculation fill the space. When the researchers filled that space themselves with a detailed letter naming their own version of events, OpenAI came back with a statement that asserts a worse offense happened than what the letter describes, without telling anyone what that offense actually was. Each side is now claiming the other is hiding the real story, and neither has produced anything a reporter can independently check. For an industry that keeps insisting it wants to be trusted with frontier AI systems, this is a strange hill to plant a credibility fight on.
None of this means the researchers are automatically right either. “Breach of trust beyond what’s outlined in their letter” is a specific enough phrase to suggest OpenAI believes it has something concrete, even if it isn’t sharing it. Companies sometimes have real legal reasons to stay vague about personnel matters, especially when sensitive technical information is involved. But reasons to stay vague and reasons to be trusted are different things, and OpenAI is currently asking for the second while only offering the first.
Related: This follows our coverage of OpenAI firing three safety researchers over a leak it still hasn’t explained, and the researchers’ open letter disputing that account, which this statement responds to.
Bottom Line: OpenAI finally said something, which is more than it managed on October 1. But “a significant breach of trust beyond what’s outlined in their letter” is a sentence built to sound damning without actually being checkable, and that’s the same problem OpenAI had a week ago, just restated in more words. If the company wants this to be the last word on the story, it needs to publish something a reporter can verify, not another line that asks us to take its word for it.
Follow Teck Hustlers: Facebook · X · Instagram · LinkedIn · TikTok · Pinterest
