Cyber Risks · Trail of Bits Blog

Trail of Bits’s Response to OSTP National Priorities for AI RFI

Summary: Trail of Bits responded to an OSTP request for information on AI policy, arguing that developers should be held accountable to verifiable, falsifiable claims that independent regulators can assess through a documented safety case, as the FDA requires in manufacturing. The firm also stated that LLMs cannot safely generate code directly but suit subtasks like suggesting changes, summarizing code, and code completion, and that AI-based vulnerability scanners using graph-based models have outperformed traditional scanners on certain vulnerabilities.

Bottom line: Trail of Bits notes that AI has significantly lowered the technical expertise and time required to carry out attacks, letting attackers rapidly develop or customize exploits against known vulnerabilities and deploy them before they are patched, which could critically affect national infrastructure.

Read the original at Trail of Bits Blog ↗