Elon Musk's big bet on "spicy" AI has officially become a financial risk factor. SpaceX disclosed the liability in its pre-IPO S-1 paperwork filed on Wednesday. The disclosure follows SpaceX's acquisition of xAI three months prior, which brought xAI's social media platform, consumer chatbot, and controversial not-safe-for-work (NSFW) Grok AI features in-house.
In the filing, SpaceX warned that such features could pose "heightened risks" and "reputational harm" because the Grok NSFW modes are "more irreverent and harsher than our standard offerings." The document flagged the potential "generation of potentially explicit content," as well as "potential nonconsensual or exploitative imagery" and intellectual property infringement. Such content could also be viewed as "harmful, harassing, abusive, or discriminatory," according to the S-1.
While flagging potential business risks—including ongoing investigations or material lawsuits—is standard practice for companies filing pre-IPO paperwork, identifying the output of generative AI as a primary corporate hazard highlights the unique regulatory challenges of modern AI integration.
Earlier this year, xAI faced severe backlash after Grok was used to generate non-consensual sexualized AI images of women, including minors. The incident sparked public condemnation from government officials, prompt policy overhauls, and legal actions. SpaceX referenced these ongoing legal challenges in its S-1, noting that the company is subject to "investigations and inquiries" in the US regarding "allegations that our AI products were used to create nonconsensual explicit images or content representing children in sexualized contexts."
"The Company and certain subsidiaries have been named as defendants in multiple lawsuits arising from Grok's image-generation and editing features," the S-1 reads, adding that SpaceX intends to "defend itself vigorously." Although Musk previously pushed back against claims regarding minor-related imagery, stating he was "not aware of any naked underage images generated by Grok," the company was forced to roll out updates restricting Grok's capabilities following the outcry.
[AgentUpdate Depth Analysis] SpaceX's inclusion of Grok’s NSFW vulnerabilities in its S-1 filing signals that AI safety and alignment have graduated from ethical debates to critical corporate liabilities. Unlike closed ecosystem leaders like OpenAI and Anthropic, which enforce strict safety guardrails on their models, xAI chose an unfiltered approach. While this strategy successfully drove viral engagement, it introduces massive compliance risks for the broader AI Agent ecosystem. As AI Agents transition from passive text generators to autonomous execution partners, the lack of robust safety baselines in foundational models like Grok could result in severe legal, financial, and operational failures. For AI Agents to achieve enterprise-scale adoption, predictable safety and transparent compliance must become non-negotiable standards.
