Anthropic whistleblower testimony airs on NDTV Profit

The most telling thing about the NDTV Profit clip on YouTube titled "Anthropic Whistleblower's Testimony on Camera" is how little the page says: a subscribe prompt and some links, with no name and no quote we could check. The story around it can be pieced together from other reporting: public departures and pointed quotes from AI-lab researchers, and a sworn hearing in New York.

If you have not been following: according to the BBC, Anthropic was formed in 2021 by a breakaway team of early OpenAI employees, positions itself as more safety-oriented and is a public benefit corporation that in 2025 agreed to pay $1.5bn to settle an authors' class action. The BBC also reported that Mrinank Sharma, who led a safeguards research team, quit in February 2026 with a letter saying "The world is in peril" and planned to study poetry.

Whose resignation is behind the funding questions?

The answer, according to ContentGrip, is Jacob Coxon, who had worked on pretraining research at Anthropic and OpenAI and announced his resignation in a lengthy X post: "Neither company is acting responsibly." ContentGrip reports that a Wall Street Journal exclusive ran alongside it, turning a personnel departure into a coordinated public announcement the company did not control.

ContentGrip, citing Deadline and Axios, says the post drew nearly 76 million views overnight and passed 115 million by the next day. Yahoo News quotes Coxon writing that "The people building AI earnestly believe that it could kill us all by the end of the decade," and reports his view that only heavy government intervention, or every major lab agreeing to slow down together, could change that. He added that neither was happening.

MindStudio, in one analysis piece, says the post has drawn questions about funding and timing coordination. ContentGrip adds a detail that complicates the simplest version of that doubt: Coxon said he left about two months before his Anthropic equity would have vested. Vesting is the schedule on which granted shares actually become yours, so leaving early is like walking out of a lease before the deposit comes back.

Leaders from Anthropic, OpenAI, Google and Meta testified under oath in New York on October 5

Separately from the video, senior leaders from Anthropic, OpenAI, Google and Meta testified under oath, on the record, about AI risks at a New York City Council hearing held on Monday, October 5. Pressure is coming from Washington as well: Sen. Richard Blumenthal has separately demanded answers.

Sen. Bernie Sanders and Rep. Greg Casar have reportedly introduced legislation to ban superintelligent AI and freeze development until safety rules exist. All of this lands as Anthropic was reportedly preparing for an IPO that could value it close to $2 trillion, which, according to the reports, helps explain the investor-facing angle of the coverage.

Who else has spoken up since the post?

People inside Anthropic have, according to ContentGrip. Researchers Evan Hubinger and Samuel Marks acknowledged that severe AI-risk concerns are genuinely held at the company, and Hubinger said Anthropic was trying its best but did not yet have a plan to solve alignment for superintelligence. Later reporting cited by ContentGrip quotes a spokesperson saying the company had long been transparent about AI's benefits and risks.

Yahoo News adds that Hubinger puts a greater-than-10% chance on AI wiping out humanity within the next decade. Explainx reports that two more researchers resigned: Joe Benton, who led a safety research team at Anthropic, and Josh Engels of Google DeepMind, who gave NBC News interviews published September 9, 2026, and are both joining METR, an independent nonprofit evaluator.

According to Explainx, Benton said transparency from the companies is "entirely voluntary," and Engels said "There are no adults in the room." Both cite the July 2026 Hugging Face breach carried out by autonomous AI agents running an unreleased OpenAI model.

The biggest gap is the clip itself: the page description gives no name and no quote, so we cannot tell whether it shows Coxon, Benton, Sharma or someone else. Coxon's forgone equity, as ContentGrip reports it, speaks to motive, not to whether his risk argument holds, and most of the backstory rests on single outlets. In our view, a title promising testimony on camera from a page that names nobody is a thin basis for the word whistleblower.

Naming the person on camera. The next fact worth waiting for is a name: until NDTV Profit or the speaker says who is testifying, the clip cannot be matched to any departure described here. A published record of the October 5 hearing would be the other checkpoint, showing what lab leaders said under oath. No vote date has been given for the Sanders and Casar bill, and the reporting we have sets no date for Anthropic's IPO.