Tag
The White House issued a memorandum directing federal agencies to create a program authorizing vetted private U.S. companies to conduct government-overseen cyber surveillance and effects operations against transnational criminal organizations engaged in cybercrime.
This paper investigates using an 8-billion parameter LLM to improve autonomous cyber defense, then distills its policy into a lightweight 64,910-parameter RL agent, demonstrating feasibility across CybORG scenarios.
This paper examines how AI-mediated civilian cyber operations challenge the direct causation criterion of the ICRC's test for direct participation in hostilities, arguing that autonomous multi-agent systems break the causal link required for classification under international humanitarian law.
Anthropic has embedded engineers within the NSA to help deploy Mythos for offensive cyber operations, raising concerns about AI involvement in surveillance.