Anthropic’s browser agent got hijacked 31.5% of the time before safeguards engaged (opens in new tab)

Across the frontier labs, the highest prompt injection figures published this spring are Anthropic’s. Point a red-teamer at in a browser, and the attacker hijacked it before safeguards engaged. OpenAI, Google, and Meta never gave security leaders a comparable number to set beside it. That figure looks like a liability. In this comparison, it is the opposite. It's the one solid piece of ground.Four frontier labs each shipped a prompt injection disclosure, and no two match. Anthropic put and fo...