Attackers who probe large language models rarely give up after one refusal. They reframe, build context across turns, adopt personas, and escalate gradually. New research from Cisco’s AI threat intelligence team finds that the safety benchmarks used across the industry miss almost all of this behavior, and the gap between published scores and observed resilience runs wide enough to misrank leading models. Single-turn versus multi-turn ASR by model, with approximate 95% confidence half-widths on single-turn … More
The post Frontier AI models collapse under multi-turn AI attacks, Cisco finds appeared first on Help Net Security.