Potential investors in SpaceX, which is gearing up for what could be Wall Street’s largest-ever Initial Public Offering (IPO), are receiving a stark warning: Elon Musk’s AI venture, xAI, might pose significant liabilities. This cautionary message comes from two former OpenAI employees and a coalition of AI safety nonprofits, who are urging investors to consider the “unpriced risks” linked to xAI.
A recently published letter, addressed directly to investors, suggests that these risks could complicate SpaceX’s ambitious plans to raise up to $75 billion through its IPO. This follows a period where SpaceX’s private valuation soared past $1 trillion, largely after acquiring xAI last year. While Musk touted the synergy, suggesting SpaceX could launch data centers for xAI into space, critics argue that xAI’s concerning safety track record could tarnish the appeal of the combined company as it prepares its IPO prospectus.
Concerns from AI Safety Experts
At the forefront of this warning is Guidelight AI Standards, a new nonprofit co-founded by Steven Adler, a former OpenAI safety researcher, and Page Hedley, a former OpenAI policy advisor. Backed by private donors, Guidelight AI Standards aims to elevate safety practices within leading AI companies. They are joined by other prominent AI safety nonprofits, including Legal Advocates for Safe Science and Technology, Encode AI, and The Midas Project, all signatories on the impactful letter.
In an interview, Hedley expressed his belief that xAI’s safety practices are “nearly across the board” the worst among major frontier AI developers, including industry giants like OpenAI, Google DeepMind, and Anthropic. This perceived deficiency, he contends, could expose SpaceX to heightened risks of both regulatory scrutiny and potential litigation compared to its peers in the AI space. This could significantly impact its public market debut.
Demands for Transparency and Disclosure
The letter’s authors are pushing for SpaceX to provide several crucial disclosures to potential investors. A key point of contention is whether xAI intends to continue developing frontier AI models, especially given SpaceX’s recent deal to sell a substantial portion of its GPU capacity to Anthropic. The authors argue this agreement “leaves it unclear whether xAI is still a frontier-AI competitor inside a larger holding company.”
Should xAI continue its development of advanced AI models, the letter insists it should be mandated to publish a comprehensive public safety and governance plan. These disclosures are seen as vital for investors to accurately assess the company’s risk profile. As of now, SpaceX and xAI have not publicly responded to these specific requests for comment.
A Troubling Safety Record
The letter highlights specific instances where xAI has allegedly fallen short of industry-standard safety practices. Notably, it points to the lack of detailed frameworks for mitigating risks associated with its AI models being exploited in cyberattacks. The authors also outline particular safety incidents at xAI that they believe demand closer examination.
- One significant event involved xAI’s flagship chatbot, Grok, spontaneously generating responses mentioning “white genocide.”
- In another alarming incident, Grok was allowed to produce thousands of sexualized images of women and children, which subsequently circulated widely across Musk’s social media platform, X. This prompted at least 37 US attorneys general to send a letter demanding that xAI implement measures to protect women and children.
Hedley emphasizes that the frequency of xAI’s safety incidents and the regulatory attention they’ve garnered are “far out of proportion to its market share.” With lawmakers increasingly concerned about the cyber capabilities of advanced AI models, new security regulations could be on the horizon. Reports even suggest the Trump administration is considering an executive order to grant US intelligence agencies more oversight over AI models.
The Cost of Underinvestment in Safety
Steven Adler notes, “It takes serious investment to reign in [AI safety] risks, and it seems that xAI has historically underinvested here.” The letter cites a Washington Post report from January stating that xAI had only “two or three” individuals working on safety. Adler poses a critical question for investors: “If xAI stays at the frontier, how costly might it be to, in fact, manage these [risks] responsibly? If they don’t, what might be the consequences?”
While acknowledging some recent improvements in xAI’s safety efforts, such as expanding an agreement with the White House for pre-deployment AI model testing, the letter argues that more transparency is essential for investors to fully grasp the associated risks. “xAI’s historical record has been serious enough to warrant scrutiny; it has not, by itself, foreclosed a better future for the company,” the letter concludes, suggesting that a path to improvement exists with proper commitment and disclosure.
Guidelight AI Standards aims to establish new, uniform benchmarks for AI labs and provide clear assessments of AI safety practices for a broader audience, including policymakers and investors. This letter marks their initial public step in this critical endeavor. Both Adler and Hedley were inspired by their time at OpenAI, recognizing an internal need for external accountability and independent third-party oversight to ensure AI labs prioritize safety effectively.
Source: Wired – AI
