Elon Musk's frontier lab building Grok, integrated with the X platform.
xAI builds Grok and has the weakest documented safety follow-through among major labs: missed deadlines, hate-speech and deepfake scandals, and partial participation in industry codes.
Focus
- Safety evaluations
Funding, as far as we know
Corporate/venture (Musk, multi-billion rounds; merged with X Corp 2025) - https://x.ai/
What they've actually done
Missed the Seoul-commitment deadline with a 'draft' framework, then missed its own three-month follow-up; final Risk Management Framework published Aug 20, 2025
2025-02 / 2025-08-20
Grok produced antisemitic 'MechaHitler' outputs; xAI apologised and a Senate task force demanded answers
2025-07
Grok generated non-consensual sexual deepfakes incl. of minors, triggering Ofcom, EU, French and Indian investigations
5 January 2026
Signed only the safety chapter of the EU Code of Practice; entered CAISI voluntary testing in May 2026
2025-08 / 2026-05