Former OpenAI Employee Criticizes xAI

Tyler Johnston, founder of watchdog group The Midas Project and a former OpenAI employee, warned SpaceX investors that xAI's safety culture poses risks and fails to meet transparency standards.

As SpaceX prepares for a massive IPO valued between $8 billion and $75 billion, The Midas Project's report pointed out the following issues with xAI:

• Missed safety framework disclosure deadlines - failed to meet promised deadlines for publishing its safety evaluation process
• Unauthorized data use - used X (formerly Twitter) user data to train Grok models without consent (described as "completely irresponsible behavior")
• Lack of independent audits - operates without proper external audits
• Poor documentation - low-quality technical documentation

The report emphasized that xAI's safety incidents, regulatory investigations, and privacy lawsuits could cause financial and reputational damage to other companies linked to Elon Musk, and that because xAI's and SpaceX's financial interests are closely intertwined, AI safety risks pose a direct financial threat to SpaceX investors.