Watchtower
Watchtower tracks when AI companies change their safety policies — and when they break them.
xAI updated their Acceptable Use Policy on Aug. 14. It renamed “xAI” to “SpaceXAI” throughout, as well as making two small changes.
xAI
Aug 14, 2026
xAI updated their Acceptable Use Policy on Aug. 14. It renamed “xAI” to “SpaceXAI” throughout, as well as making two small changes.
Instead of just prohibiting “Using the Service or any Output to develop machine learning models or related AI services that compete with xAI,” the new policy also prohibits using it to develop “any products or services that compete with SpaceXAI, whether directly or indirectly.”
In the previous version, SpaceXAI prohibited “Accessing the Services through automated or non-human means, whether through a bot, script, or otherwise,” the updated AUP only prohibits accessing Services “through unauthorized automated or non-human means.”
Oct 13, 2026
Google launched Gemini 4 Argon along with the first “full” safety and preparedness report made under its Frontier Safety Framework (v3.1), published in April 2026.
Meta
Oct 12, 2026
Meta publicly launched Muse Wonder with a 13-page model card and without concurrently publishing a safety and preparedness report.
Meta
Oct 2, 2026
Meta updates and renames its safety framework from the "Advanced AI Scaling Framework" (v2) to the "Meta Superintelligence Scaling Framework" (v2.1).
OpenAI
Sep 29, 2026
OpenAI updated the GPT-6 Astra system card to include information about Dots, an always-on agent powered by GPT-6 Astra.
OpenAI
Sep 22, 2026
OpenAI Updated the GPT-6 Astra system card to include information about GPT-6 Sol and GPT-6 Luna. There were also updated alignment evaluations and new results for Astra.
Anthropic
Sep 22, 2026
Anthropic’s release of Opus 5.5 with an “inconclusive” determination for harmful-manipulation.
Frequently asked questions
Have more questions? Our team is happy to help, contact us.