OpenAI’s Safety and Futures Chief Joshua Achiam Is Departing After Nearly Nine Years

OpenAI is losing one of the executives most closely associated with its public-facing safety mission. Joshua Achiam, the company’s chief futurist and a longtime internal advocate for alignment and policy, told colleagues this week that he will leave later this month after nearly nine years at the company.

His exit arrives at a moment when OpenAI is trying to scale rapidly, formalize its governance, and prepare for an eventual public-market debut. It also underscores a broader pattern: as the AI industry matures, the people who spent years arguing for caution, restraint, and long-term risk management are increasingly moving on, often to roles outside the frontier-lab model they once helped build.

Achiam said his decision was not prompted by a single event. Instead, he framed it as something he had been considering for some time, and suggested that the broader conversation around advanced AI has shifted enough that the work can now be advanced beyond OpenAI’s walls.

The world is in on the secret now and it feels possible to work on the mission from outside the walls of a frontier lab, Achiam told staff in a note seen by WIRED. He added that he believes humanity can reach a future of peace, remarkable prosperity, and extraordinary scientific and social progress, and said he intends to keep contributing to that goal in whatever comes next.

OpenAI has not said whether it will replace Achiam in the role, which sat at the intersection of safety, policy, and long-range strategic planning. The company’s lack of an immediate successor points to how fluid these responsibilities have become inside a fast-changing organization that has repeatedly reshaped its research and governance structure since ChatGPT’s launch in 2022.

Why Achiam mattered inside OpenAI

Achiam’s job was not a conventional product or engineering position. As chief futurist, he occupied a niche but important space: thinking about how the next waves of AI development might affect users, competitors, regulators, and society at large. That included evaluating possible harms, identifying policy priorities, and helping senior leadership think through the long-range implications of highly capable models.

He worked closely with OpenAI’s Global Affairs chief, Chris Lehane, to argue for regulatory approaches that would be compatible with the company’s stated mission of ensuring that artificial general intelligence benefits all of humanity. In practical terms, that meant connecting technical research to government relations and public policy, a linkage that has become central to nearly every major AI company as the stakes around model deployment have intensified.

At OpenAI, Achiam was known internally as a strong voice for safety. He also had a reputation for being willing to push back, including on the broader AI safety community when he believed the discussion had become too detached from the realities of product development and deployment.

A role that evolved as OpenAI scaled

OpenAI’s internal structure has changed repeatedly in the three years since ChatGPT turned the company into one of the most influential technology firms in the world. What began as a relatively compact research organization became, in short order, a sprawling business with consumer products, enterprise offerings, infrastructure demands, and an expanding political footprint.

That growth has forced OpenAI to repeatedly redraw the boundaries between research, safety, product, and policy. The company has reorganized those functions several times, sometimes to speed up decision-making and sometimes to better integrate technical and governance work.

In 2024, OpenAI announced a “mission alignment team” led by Achiam, a signal that the company still wanted an internal mechanism focused on keeping its work tied to its founding purpose. By February, that team had been dissolved, and Achiam was reassigned to a newly defined role as chief futurist. His departure now closes another chapter in OpenAI’s ongoing effort to balance growth with caution.

What the mission alignment team represented

The team was meant to serve as a kind of internal compass, helping the company evaluate whether new products and research directions matched OpenAI’s broader obligations. Its existence reflected a persistent tension in the company’s identity: it was at once a commercial business shipping products at scale and a research organization with origins in a nonprofit ideal.

Its dissolution suggests that OpenAI has been consolidating those responsibilities into other parts of the company, particularly as policy and safety leaders increasingly work directly with researchers rather than through separate mission-focused structures.

OpenAI’s safety leaders keep leaving

Achiam is the latest in a series of departures among OpenAI employees most associated with AI safety, governance, and long-horizon risk analysis. Those exits have become more notable as the company races ahead with product launches, partnerships, and capital formation.

Some former leaders have gone to competing AI labs. Others have launched new organizations that aim to influence the field from the outside. The pattern suggests that, for many in the safety community, the center of gravity is no longer exclusively inside OpenAI.

Among the more prominent departures:

• Jan Leike, who helped lead OpenAI’s Superalignment initiative, left in 2024 for Anthropic.
• Miles Brundage, OpenAI’s former head of policy research, departed in 2024 to create a nonprofit focused on strong AI safety and security norms.
• Steven Adler, who directed research into dangerous AI capabilities, also left in 2024 to found a nonprofit advocacy group.
• Andrea Vallone, who worked on how ChatGPT should respond to users in emotional or mental distress, later joined Leike’s team at Anthropic.

The cumulative effect is more than a list of personnel changes. It reflects a redistribution of expertise across the AI ecosystem, with safety researchers increasingly splitting between labs, advocacy groups, and independent policy efforts.

Dean Ball arrives as OpenAI retools its future-facing work

Even as Achiam exits, OpenAI is adding another figure with policy and government experience. Dean Ball, a former White House AI adviser, joined the company this week as head of strategic futures. The overlap between Ball and Achiam will be brief, according to the reporting, but the move indicates that OpenAI is still investing in people who can bridge technical forecasting and policy development.

Ball is expected to work closely with researchers and policy staff. That makes his role especially relevant at a time when OpenAI is trying to anticipate what kinds of regulations, safety frameworks, and standards could shape the next phase of AI deployment.

There is a strategic logic to that hiring. As models become more capable and more widely embedded in consumer and enterprise products, companies need staff who can think several moves ahead: not just about what the technology can do today, but what lawmakers, regulators, and the public may demand tomorrow.

The broader context: OpenAI is changing as it prepares for public markets

Achiam’s departure also arrives as OpenAI moves closer to a more conventional corporate future. The company has been widely expected to pursue a public listing or a structure that more closely resembles one, a shift that would increase pressure on leadership to deliver growth, stability, and predictable governance.

That transition is relevant because safety work can become harder to prioritize when companies face intensified scrutiny from investors, partners, customers, and rivals. The same is true when organizational attention shifts toward commercialization and infrastructure scaling.

OpenAI’s mission language remains central to its branding, but the practical demands of operating one of the world’s most important AI companies can create friction with the slower, less visible work of anticipating societal risk. The departure of a chief futurist who helped formalize that mission-sensitive thinking raises questions about how OpenAI will preserve institutional focus on long-term safety as it expands.

Why safety leaders are moving out

There are several reasons why safety and policy leaders may be leaving AI labs:

1. The work may feel more impactful in independent organizations or government-adjacent roles.
2. Internal safety teams can become less central as product development accelerates.
3. Researchers may believe that external pressure, regulation, and public accountability can shape the industry more effectively than internal advocacy alone.
4. As labs grow, decision-making can become more centralized and less receptive to dissent.

None of those explanations applies identically to every departure. But taken together, they help explain why a number of influential AI figures now see better leverage outside a single company, even one as powerful as OpenAI.

Achiam’s path from intern to chief futurist

Achiam’s long tenure at OpenAI began in 2017, when he joined as an intern. Over time, he rose into research, eventually becoming a research scientist focused on AI safety. That progression gave him a front-row view of OpenAI’s transformation from a compact, mission-driven lab into a dominant force in the commercial AI market.

Inside the company, he earned a reputation as a steadfast defender of the safety mission. At the same time, he was not shy about challenging the conventional wisdom of parts of the AI alignment world, a stance that made him a distinctive voice in a field often defined by sharp internal debates.

His departure is therefore not just a personnel matter. It marks the exit of someone who had been present for almost the entire arc of OpenAI’s rise, and who helped shape how the company talked about its responsibilities to the public.

The Elon Musk episode that followed him into court

Achiam also became known beyond OpenAI for his role in a notorious early company moment involving Elon Musk. In testimony earlier this year, Achiam said he had interrupted Musk during the billionaire’s farewell remarks when Musk was leaving OpenAI in 2018. Achiam reportedly objected to Musk’s plans to pursue AGI development through Tesla, warning that safety could be compromised if the work moved too quickly.

According to the testimony, Musk responded by insulting him. The exchange later became a kind of inside legend in AI circles, particularly after Dario Amodei — now chief executive of Anthropic — and David Luan, who later became head of Amazon’s AGI lab, presented Achiam with a humorous trophy: a golden donkey’s rear end engraved with a message praising his willingness to be, in effect, difficult in the service of safety.

That story has circulated as a symbol of a deeper divide in AI: whether being cautious is a noble and necessary discipline, or an obstacle to progress. Achiam’s career suggests that, in practice, the answer is often both depending on who is doing the judging.

What his exit says about the AI safety debate

The departure of a high-profile futurist from OpenAI lands in the middle of a larger debate about where AI safety work belongs. For years, the assumption among many researchers was that the most important decisions would be made inside the companies building the most powerful models. But as frontier AI becomes more systematized, some of those same experts now argue that regulation, standards, and public oversight may matter just as much as internal controls.

Achiam’s own wording reflects that shift. Rather than framing his exit as a withdrawal from the mission, he described it as a change in venue. That is a subtle but meaningful distinction. It suggests he believes the field has matured enough that the work can continue through advocacy, policy, or external collaboration rather than direct participation in a single lab’s hierarchy.

For OpenAI, that creates both a challenge and an opportunity. On one hand, losing experienced safety leaders can weaken institutional memory and complicate the company’s ability to speak credibly about governance. On the other, the company may see value in distributing expertise across the broader ecosystem, where it can influence standards in multiple directions.

OpenAI’s balancing act going forward

OpenAI is now trying to manage several competing imperatives at once: expand commercially, remain technically competitive, reassure regulators, and preserve a public image rooted in responsibility. That is a difficult equilibrium for any company, and particularly for one that has come to symbolize both the promise and the anxiety of generative AI.

Personnel changes matter in that environment because they reveal where an organization is placing its attention. Hiring a strategic futures leader while losing a chief futurist could be read as continuity or as a reshuffling of priorities, depending on how OpenAI defines the new role and whether it gives it real influence.

What is clear is that the company is still trying to build a policy apparatus sophisticated enough to keep pace with its research output. With major model releases, enterprise adoption, and potential regulatory headwinds all in play, the next set of decisions will likely be shaped as much by governance as by engineering.

Key facts at a glance

What to watch next

There are several questions to watch in the wake of Achiam’s exit.

• Whether OpenAI names a direct replacement for the chief futurist role.
• How much influence Dean Ball and other policy-adjacent hires will have over model development.
• Whether additional safety-focused employees leave as the company continues to scale.
• How OpenAI explains the relationship between its mission language and its commercial strategy.
• Whether the company expands or retools its internal approach to long-term risk assessment.

For now, Achiam’s departure is a reminder that AI governance is no longer just an internal debate among researchers. It is becoming a broader institutional contest over who gets to shape the future of the technology — companies, governments, nonprofits, or some combination of all three.

And if Achiam’s message to colleagues is any indication, he expects to remain part of that contest, just from a different vantage point.