How DeepMind Red‑Teams AI—and Why It Matters

Two of Google DeepMind’s leaders—Dawn Bloxwich (Responsible Development & Innovation) and Tom Lue (VP, Frontier AI Global Affairs)—open the playbook on frontier safety and policy. Bloxwich explains how DeepMind blends structured evaluations with red‑teaming by experts and jailbreakers, and how its Frontier Safety Framework addresses severe risks (from CBRN and cyber to loss of control and socio‑affective concerns). Lue maps the regulatory landscape across the EU, US and Asia, arguing for harmonized standards that safeguard without stifling innovation. Together, they show what responsible development looks like before and after deployment—model cards, third‑party testing and the industry forums trying to align best practices.

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.