Ex-OpenAI staff warn SpaceX IPO investors about xAI safety risks

In short: Two former OpenAI employees and several AI safety nonprofits sent a letter warning that xAI’s safety issues could create risks for SpaceX investors ahead of a possible IPO.

What happened

Two former OpenAI employees, Steven Adler and Page Hedley, joined several nonprofits to publish an open letter aimed at prospective SpaceX investors. The letter says SpaceX should share more information about xAI, Elon Musk’s AI lab, before SpaceX goes public. Going public is when a company starts selling shares on the stock market.

The letter argues that xAI could bring “unpriced risks” to SpaceX after SpaceX acquired xAI last year, according to WIRED. The authors say these risks could affect how investors evaluate SpaceX as it prepares an IPO prospectus, which is the formal document that explains a company’s finances and major risks.

The group behind the letter includes a new nonprofit called Guidelight AI Standards, plus other AI safety organizations. Hedley told WIRED he believes xAI’s safety practices are weaker than those of other major AI developers.

The letter points to past incidents involving xAI’s chatbot Grok, including responses that mentioned “white genocide” and the generation of sexualized images that spread on X. The letter also cites reporting that xAI had only “two or three” people focused on safety as of January.

The authors also ask SpaceX to clarify whether xAI will keep building “frontier” AI models, meaning the most advanced models (like the newest, most powerful engines). They note SpaceX recently made a deal involving GPU capacity, which is high-powered computer chips often used to run AI.

SpaceX and xAI did not immediately respond to WIRED.

Why it matters

For everyday investors, this is about hidden costs and legal risk. If regulators step in, or lawsuits follow, it could affect SpaceX’s business and the value of its stock.

Source: Wired