Comment by Stephen Casper

AI safety researcher at the UK AI Security Institute, researching safeguards for open-weight AI models.
Open models have a tendency to be less safe a lot of the time than closed ones. Maybe it could be misused by someone who downloads it and modifies it to remove safeguards or to sharpen its harmful capabilities. This type of fine-tuning could cost hundreds or a few thousand dollars, at most.
AI Verified (Oct 19, 2026)
Like Share on X 21d ago

Policy proposals and claims

votes For
Statement relation verification history AI Verified Report this

Statement relation comments

AI Verified Casper explicitly describes open models as less safe than closed ones and identifies harmful misuse through removal of safeguards. · Hector Perez Arenas gpt-5 · 21d ago
Vote inference verification history AI Verified Report this

Vote answer comments

AI Verified Describing open models as less safe and susceptible to harmful safeguard-removing fine-tuning supports a for vote. · Hector Perez Arenas gpt-5 · 21d ago

Quote authenticity verification history

Report this

Quote authenticity comments

AI Verified Harvard Gazette’s 19 October 2026 interview records Casper’s statement about open models, safeguards, and harmful fine-tuning. · Hector Perez Arenas gpt-5 · 21d ago
replying to Stephen Casper

Other authors to follow