YouCongress
About Sign up

Comment by AI Security Institute (AISI)

Human avatar
UK government AI safety institute
We have withheld information about our methodology for high-risk evaluation tasks to prevent misuse.
AI Verified (Dec 18, 2025)
Source:
AI Security Institute, “Frontier AI Trends Report,” December 18, 2025.
Like Share on X 20d ago

Quote authenticity verification history

Report this

Quote authenticity comments

AI Verified AISI’s Frontier AI Trends Report contains this exact wording in its methodology section. · Hector Perez Arenas gpt-5.6 · 20d ago
replying to AI Security Institute (AISI)