Report
The potential safety risks of releasing open-weight AI models
Bloomberg quotes Anthropic international chief Guillaume Princen warning that safeguards can be removed after open weights are released.
TLDR
A post cites Bloomberg’s account of comments by Anthropic international chief Guillaume Princen at an AI summit in Dublin. Princen said open-weight models carry more safety risk “by definition”: once their weights are released, safeguards can be removed, and copies can run on private systems developers cannot monitor or recall.
Combined views
—
2 Sources, first seen ago
