Claude users found ways around safeguards for bioweapons research

Some dangerous biology looks much like legitimate research, complicating AI safeguards.

This news item is fetched from a public RSS feed. Title, summary, and image are © their respective publisher; follow the link above to read the full article on the original site.