1 hour ago
1
Another day, another doomsday AI post that leaves the public feeling unsettled, anxious, and unsure what to do next. At least that’s how it feels as an AI reporter who reports on this space every day. This time, the alarm bells are coming from inside the industry, courtesy of a former researcher at OpenAI and Anthropic named Jacob Coxon. His X post on Sept. 8 has gone mega-viral, racking up over 150 million views as of this writing.
“The people building AI earnestly believe that it could kill us all by the end of the decade,” Coxon said. “I spent the last three years doing pre-training research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives.”
Anthropic’s current head of alignment re-posted Coxon, noting that he is “correct” that “we really do earnestly believe AI could kill all humans!” He estimates there’s a greater than 10% chance it could happen within the next decade. And earlier this week, OpenAI’s head of research Jakob Pachocki published a blog post outlining his concerns for humanity.
Not great. So what do we do about it?
This is the first time an AI warning has truly gone mainstream
While Coxon is far from the first to warn about the perils of AI, his message might be the first to truly break into the mainstream. He promptly went on a media tour the same day he posted on X, and his message captured headlines at top news organizations. Even country singer Sheryl Crow urged her Instagram followers to take him seriously.
There are a few theories as to why Coxon’s post has spread further than previous warnings about AI. As my editor Jeremy Kahn and I discussed earlier today, anxiety about AI has reached all-time highs following recent revelations about OpenAI’s AI agents escaping their sandbox and hacking the Hugging Face website. That incident provided a tangible example of how AI can go rogue and do things that would be considered a crime if a human were involved. The public may also be more ready to hear the message right now, at a time of growing concerns about energy-sucking AI data centers and about AI’s future impact on jobs.
I also think Coxon’s authoritative confidence in the existential threat posed by AI—and the quick response by Anthropic’s head of alignment agreeing that, yes, AI could very well kill us all—underscored the gravity and the urgency of the situation.
In a first for me, Coxon’s message reached my group chat with other reporters, none of whom cover technology or AI. “Emily, any thoughts on the impending AI apocalypse and what to do about it?” my friend asked me in the group. She’s an arts and culture reporter who works in local news. It struck me as not only a good question, but the question everyone should be asking right now.
Without more information, and receipts, Coxon is more likely to create anxiety than change
The problem with Coxon’s post is that it’s vague, with no proof, and no specific examples that are easy for the public—and regulators—to act on. Yet his media tour suggests he wants his message to reach a generalized audience.
Hear me out: He claims the AI could kill us someday, but doesn’t point to any projects in the pipeline that could be shut down to avoid this. He says AI companies are moving too fast, but neglects to share screenshots, emails, or specific examples of when this behavior went sideways—when it became clear to leaders at AI companies that the technology was slipping beyond their control, for instance, and how the decision-makers disregarded the warning signs. He doesn’t suggest any new legislation, name problematic leaders that should step down, or post an in-depth look at how Anthropic researches new models and propose a new approach.
To quote another viral social media clip of Heather Gay from a 2024 episode of The Real Housewives of Salt Lake City (and at the risk of seeming glib), we need “Receipts, proof, timelines, screenshots, f*cking everything!”
That’s why tech journalist Taylor Lorenz called out Coxon for “vagueposting and fomenting fear,” noting that he offered no “actual proof and receipts showing specific instances of that negligence so that it can be corrected and so that we know what you’re talking about.” As a result, she said, the generalities in Coxon’s post will only ratchet up public fear and anxiety and lead to “terrible policy.”
Ian Krietzberg, an AI correspondent at Puck News, agreed. Coxon is “not blowing the whistle on either OpenAI or Anthropic. He’s not revealing non-public information about their practices that he finds so concerning,” Krietzberg said. “The whole thing is broad ideas-based, not specific company wrongdoing.”
So what should we do? What should I tell my arts reporter friend? Coxon doesn’t say. His main call to action is geared toward his fellow AI researchers: “If you are a lab researcher, I urge you to consider what the next few years will actually feel like. Do you want to kick off a superintelligent RL run without a rigorous understanding of its mind? Should you put your head down because “it’s happening anyway” – or take this moment to call for different conditions?”
Others have taken to social media to say people should call their legislators in response to Coxon’s post, and urge more people working in AI to speak out.
Coxon’s message still has value—just not as much as it could
There are a few things I like about Coxon’s post. The main one is that he gave up his shares in Anthropic stock upon resignation, Axios reports. That shows he’s putting his money where his mouth is. And it shields him from any criticism that he stands to make millions of dollars from Anthropic’s imminent IPO while preaching that others should not do the same.
The second thing is that he provides an opening for other researchers to say they are uncomfortable with what’s happening behind closed doors. Perhaps Coxon being light on details sets an easier precedent for others, who may not want to feel the pressure of bearing the burden of proof.
That would be a worthwhile outcome—or at least, a move in the right direction to a worthwhile outcome. I don’t share the absolutist view of the critics who accuse Coxon of doing more harm than good. This is an incredibly serious issue, and we need to hear from credible voices on the inside.
At the same time, I think we would do well to accept more specificity from these researchers, and push them to elevate the dialogue in a way that can lead to actionable change. If Coxon had been more specific, we might have already seen some.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
This story was originally featured on Fortune.com
