Watchtower
Watchtower tracks when AI companies change their safety policies — and when they break them.
Quietly changed "core values" to de-emphasize social impact and emphasize an intense focus on AGI.
OpenAI
Oct 12, 2023
On October 12, 2023, Semafor reported that OpenAI had quietly changed the "core values" it lists for the company on their recruiting page. Previously, it had included values like "impact-driven," meaning that employees "care deeply about real-world implications," and replaced them with a greater focus on building AGI quickly: adding "AGI focus" and "intense and scrappy" to their values.
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).
Anthropic
Sep 22, 2026
Anthropic’s release of Opus 5.5 with an “inconclusive” determination for harmful-manipulation.
xAI
Sep 21, 2026
xAI says Grok 4.7 scores below its safety framework’s capability thresholds on dual-use knowledge, but xAI’s safety framework doesn’t provide thresholds.
OpenAI
Sep 9, 2026
Updated the GPT-6 Astra system card, specifically to hedge claims about misalignment.
OpenAI
Aug 18, 2026
OpenAI updated its Model Spec, the document outlining intended model behavior
xAI
Aug 17, 2026
xAI updated Grok 4.6's model card after release. A changelog was included.
Frequently asked questions
Have more questions? Our team is happy to help, contact us.