YouCongress
Explore About Sign up

Comment by Alan Chan

Avatar
Research Fellow at the Centre for the Governance of AI (GovAI), researching AI policy, AI agents, transparency, and technical AI governance.
“[Models] haven’t necessarily gone through a bunch of safety testing,” and “[their] internal safeguards have not been deployed.” Running with “cyber safeguards off” and “not doing enough red teaming” was “potentially a factor in some of the recent incidents.”
AI Verified (Oct 3, 2026)
Source:
Alan Chan, Research Fellow at GovAI, quoted in “We can’t trust them completely”: AI research fellows warn that labs are running models with the safeguards off behind closed doors, Alpha Leaders, October 3, 2026. The report says Chan told a Washington briefing that models inside labs, before anyone outside sees them, have not necessarily undergone much safety testing and that internal safeguards have not been deployed; it attributes his statements that cyber safeguards being off and insufficient red-teaming could have been a factor in recent incidents.
Like Share on X 1h ago

Quote authenticity verification history

Report this

Quote authenticity comments

AI Verified Source text and contemporaneous search result attribute the stored, bracketed excerpts to Alan Chan in the 3 Oct 2026 Alpha Leaders report. · Hector Perez Arenas gpt-5.2 · 1h ago
replying to Alan Chan