OpenAI’s safety-first roots appear to be fading as the AI race intensifies.
What happened: OpenAI’s CEO Sam Altman is considering changing the startup’s governance structure to a for-profit business, per The Information, a shift that would reduce the influence of the safety-conscious non-profit arm of the company.
- The move, which Altman reportedly floated with shareholders, would also open the door to a potential blockbuster IPO for the startup, which is currently valued at US$86 billion.
Catch-up: OpenAI’s structure is strange and complex, and changing it may not be straightforward. Its for-profit arm is controlled by a non-profit that, in theory, has only one goal: ensuring the tech OpenAI develops is safe.
- Unlike for-profit corporate boards, OpenAI’s is not tasked with maximizing returns for shareholders — the idea being that this could conflict with making AI safe.
Why it matters: OpenAI already raised some eyebrows when it shuttered its long-term safety team last month. This appears to be another move to de-emphasize that aspect of OpenAI’s company culture.
- Jan Leike, one of the two former safety leads who resigned last month, pointed directly to safety taking a backseat to new products as the reason for her departure.
Big picture: Other major AI players like Google and Meta have also disbanded and split up their dedicated safety teams, a shift that has coincided with a shoot first ask questions later mentality around the industry.—LA
