SecurityAnthropic expands frontier-risk testing beyond cyber and biologyAnthropic’s Frontier Red Team introduced evaluations for tactical geolocation and conventional-weapons tasks, including drone control under simulated feedback. The company says frontier models completed some work historically limited to scarce experts and uses the results to justify stronger misuse controls. These are Anthropic-designed risk evaluations, not evidence of real-world weapon performance.Anthropic · Sep 10, 2026
SecurityA provider report tracks cyber, fraud, influence, and weapons misuseAnthropic’s September threat report describes operations it says it disrupted between December 2025 and August 2026 across cyberattacks, surveillance, influence campaigns, fraud, biological misuse, weapons work, and model distillation. It offers unusually concrete misuse cases, but the evidence and attribution are published by the provider that detected and blocked them.Anthropic · Sep 10, 2026