Joshua Achiam, OpenAI's chief futurist and one of its longest-tenured safety advocates, is leaving the company later this month after nearly nine years. He told colleagues the exit wasn't triggered by any single incident — more of a slow-burn realization that the work can continue from outside the walls of a frontier lab. Whether you believe that framing or not probably depends on how closely you've been watching OpenAI's internal churn over the past two years.

In a note to staff obtained by WIRED, Achiam offered a characteristically optimistic sendoff: "The world is in on the secret now," he wrote, gesturing at a future of "peace, unprecedented prosperity, and unimaginable possibilities." Inspiring, sure. But it's also worth noting the pattern here: when safety-focused leaders leave OpenAI, they tend to write beautiful farewell notes and then land at organizations explicitly built to counterbalance OpenAI's influence. We'll see where Achiam ends up.

Who Was Achiam, Actually?

Achiam joined OpenAI as an intern in 2017 — back when the organization was a genuine research lab and not a trillion-dollar-adjacent tech company in the middle of an IPO filing. He rose to become a research scientist focused on AI safety, then led something called the "mission alignment team" starting in 2024, which was tasked with keeping OpenAI honest about its own stated purpose: ensuring that artificial general intelligence benefits all of humanity, not just the shareholders.

That team was quietly disbanded in February. Achiam was rebranded as "chief futurist" — a title that sounds like a promotion but is also, let's be honest, the kind of role organizations create when they're not sure what to do with someone. He collaborated with OpenAI's global affairs chief Chris Lehane and worked at the intersection of the company's AI safety and policy teams. OpenAI has not announced whether anyone will fill the role going forward.

The Pattern That Should Concern You

Achiam is not a one-off departure. He's the latest entry in a growing ledger of safety-focused exits from OpenAI — and the timing, right as the company is preparing to go public, raises legitimate questions about organizational priorities.

• Jan Leike, who co-led the Superalignment team focused on keeping advanced AI under human control, left for Anthropic in 2024.
• Miles Brundage (head of policy research) and Steven Adler (dangerous capabilities research lead) both departed to found safety-focused nonprofits that year.
• Andrea Vallone, who led research on how ChatGPT should handle users in mental or emotional distress, followed Leike to Anthropic at the end of 2025.

Each of these exits came with thoughtful farewell notes and optimistic framing. The cumulative effect, though, is an organization that keeps losing the people whose job was to pump the brakes — right as it's accelerating hardest.

Replacing One Futurist With Another

Former White House AI adviser Dean Ball started at OpenAI this week as head of strategic futures, briefly overlapping with Achiam. Ball is expected to work across research and policy — roughly the same mandate Achiam held. Whether his background in government advising translates into the kind of internal safety advocacy Achiam represented remains to be seen. The title is adjacent; the institutional knowledge and credibility accumulated over nine years is not transferable.

Meanwhile, OpenAI has been making a deliberate push to bridge its research and policy teams, with researchers like Boaz Barak, Noam Brown, and Adrien Ecoffet reportedly getting more involved in policy discussions. That's a meaningful structural change — if it sticks, it could partially compensate for losing dedicated safety-policy liaisons like Achiam. Or it could mean researchers are getting pulled into political work while the actual technical safety research suffers. Both outcomes are plausible from the outside.

The Golden Donkey Footnote

No writeup of Achiam would be complete without the anecdote that turned him into something of a folk hero in AI safety circles. Earlier this year, he testified in federal court that he interrupted Elon Musk's farewell speech at OpenAI in 2018 to argue that Musk's plan to develop AGI at Tesla could compromise safety standards. Musk allegedly called him a "jackass" for the interruption.

In response, Dario Amodei — now CEO of Anthropic — and David Luan — now heading Amazon's AGI lab — gifted Achiam a golden statue of a donkey's rear end, inscribed with: "Never stop being a jackass for safety." It's the kind of story that writes itself, and it tells you something real about Achiam's role at OpenAI: he was the guy willing to say the uncomfortable thing in the room, even to the most powerful person in it.

The question OpenAI now has to answer — probably without any ceremony — is who plays that role next.