OpenAI has quietly disbanded its preparedness team, the unit tasked with assessing long-term risks from advanced AI systems. The move comes as the company reshuffles its internal structure ahead of an expected initial public offering.
Why the team was dissolved
The preparedness team was folded into other parts of the organization during a broader restructuring. OpenAI has not said what prompted the change, but the timing aligns with its reported plans to go public. The company is believed to be streamlining operations to present a cleaner, more investor-friendly picture.
That streamlining appears to have come at the cost of a dedicated safety unit. The team was responsible for evaluating catastrophic risks, including biosecurity threats and AI's ability to self-replicate. Its dissolution doesn't mean those functions vanish entirely, but they're now spread across different departments with competing priorities.
Safety concerns take a back seat?
Researchers and former staff have privately voiced worries that breaking up the preparedness team will dilute focus on safety. The concern is that with an IPO on the horizon, risk management becomes secondary to hitting financial targets. OpenAI has historically positioned itself as a safety-first lab, but its recent trajectory suggests a shift toward commercialization.
There's no suggestion that the company is abandoning safety work altogether. But the optics are uncomfortable. Disbanding the exact group that was supposed to catch worst-case scenarios right before Wall Street gets involved doesn't inspire confidence.
Risk management under the microscope
The IPO process will force OpenAI to disclose its risk factors in detail. That's where the real scrutiny will land. Regulators and potential investors will want to know how the company handles the dangers its own researchers flagged. If the preparedness team's dissolution is seen as a downgrade of those concerns, it could become a liability.
OpenAI isn't commenting beyond confirming the structural change. That silence leaves a gap. The company has yet to explain how it plans to maintain the same level of oversight without a dedicated team.
The restructured setup may work fine, but it's a gamble. AI safety isn't a problem that gets solved by spreading responsibility thin. It needs focused attention, and right now OpenAI's attention is split between building products, raising capital, and preparing for a public debut.
No date has been set for the IPO, but reports suggest it could come within the next year. Before that happens, OpenAI will need to file paperwork that reveals more about its internal governance. That filing could shed light on whether the preparedness team's functions were truly absorbed or quietly shelved.
Until then, the question hangs in the air: can a company racing to go public still take existential risk seriously? OpenAI's next moves will answer that.
