SpaceX warned investors that AI features such as Grok’s “Enriched” and “Unenhanced” modes, which allow the chatbot to produce pornographic images or voice responses with few security filters, could expose the company to regulatory scrutiny and reputational damage, according to a filing submitted Wednesday as part of the company’s planned initial public offering.
As of December, SpaceX had set aside $530 million for possible losses from lawsuits, some of which may be due to ongoing complaints filed against its AI division about sexual images generated by the Grok chatbot.
The disclosure shows how SpaceX took new financial and reputational risks when it acquired Elon Musk’s artificial intelligence startup xAI in February, a deal that sent the rocket maker’s private valuation soaring above $1 billion. In the filing, SpaceX reiterates that xAI’s goal is to develop “truth-seeking artificial intelligence.” In practice, that often meant introducing AI features with minimal guardrails. While Grok’s free-wheeling nature is often held up by Musk as a selling point, it has landed xAI in hot water with regulators.
Disclosing potential business risks is a necessary and legal part of an IPO filing, and some of the concerns raised by SpaceX may not materialize. The company is one of the chatbot makers under scrutiny by regulators as governments grapple with the societal implications of artificial intelligence tools.
SpaceX disclosed in the filing that it is currently under investigation in the United States and other countries over allegations that Grok was used to create child pornography. The company also noted that it is a defendant in several ongoing class-action lawsuits, and that “misuse” of its AI products in the future could result in additional sanctions, “including the loss of access to certain markets, which has occurred in the past.”
Some of SpaceX’s AI products, including Grok’s Spicy and Unhinged methods, “are designed to produce results that are clear, precise, or slightly reserved or rude,” the filing notes. “Because these methods may be more disrespectful and cruel than the usual offers, they present a higher risk, including damage to reputation, the production of content that may be transparent and false facts or misleading results, possible images of disagreement or abuse, infringement of the law of inventions, or content that may be considered exploitative, abusive, abusive, harassing.”
SpaceX also disclosed to investors that Grok and X had about 550 million combined monthly users as of March 31, according to the filing. Of those, 117 million use Grok’s AI features each month. In comparison, OpenAI says ChatGPT has more than 900 million weekly users.
Whether the risks posed by Grok and X are worth the headache may be one of the key questions investors will have to grapple with ahead of the SpaceX IPO. Earlier this week, a group of nonprofits warned that xAI’s poor safety record could be a liability for SpaceX investors.
SpaceX’s AI unit, which includes iX and xAI, is dragging down the entire company, with an operating loss of more than $6.3 billion last year. Sales of ads, data, and subscriptions are growing, but not at a pace that would make segmentation immediately profitable. One bright spot for SpaceX’s AI efforts is its deal with Anthropic, which has agreed to pay $15 billion a year for access to the company’s data centers.
