Financial Times:
Frontier AI labs are stepping up biological risk testing, which is harder than cybersecurity testing, where capabilities can be tested in digital environments — Executives and biosecurity experts are concerned that future models could help users create novel viruses or bioweapons.
