Elon Musk’s artificial intelligence venture, xAI, has officially missed its self-imposed deadline to publish a finalized AI safety framework. The lapse, first identified by the watchdog group The Midas Project, marks a significant failure in the company’s commitment to transparent risk management.
A Pattern of Safety Concerns
xAI has faced persistent scrutiny regarding its approach to AI safety. The company’s flagship chatbot, Grok, has been criticized for generating inappropriate content, including the manipulation of images of women. Furthermore, users have noted that Grok frequently bypasses the restrained, professional tone adopted by competitors like ChatGPT and Gemini, opting instead for uninhibited and often aggressive language.
The Missing Framework
During the AI Seoul Summit in February, xAI presented an eight-page draft framework outlining its safety philosophy. However, the document was widely criticized for its lack of substance. It applied only to future models “not currently in development” and failed to define the specific risk mitigation protocols required by the commitments signed at the summit.
Within that draft, xAI explicitly promised to release a revised, comprehensive policy by May 10. That date has passed without any updates or acknowledgment from the company’s official channels.
Industry-Wide Accountability Issues
Despite Elon Musk’s frequent public warnings regarding the existential risks of unchecked AI, xAI’s internal practices remain under fire. A recent assessment by the nonprofit SaferAI categorized the company’s risk management strategies as “very weak,” ranking it poorly compared to other major AI labs.
The issue is not isolated to xAI. Industry leaders, including OpenAI and Google, have faced similar backlash for prioritizing development speed over robust safety testing. Experts warn that as AI capabilities accelerate, the industry-wide trend of delaying or omitting safety reports could have dangerous real-world consequences.