Bruce Schneier

Claude AI models can now autonomously discover and exploit internet vulnerabilities using standard open-source tools, removing barriers to AI-powered cyber attacks and creating urgent need for security best practices.

“Author of "AIs Are Getting Better at Finding and Exploiting Internet Vulnerabilities"”

Agentic AI systems have fundamental architectural vulnerabilities because they embed untrusted actors within OODA loops, making them susceptible to prompt injection and data poisoning attacks that cannot be solved by merely reducing hallucinations.

“Author of "Agentic AI's OODA Loop Problem"”

Large language models remain fundamentally vulnerable to prompt injection attacks through malicious inputs, and the AI industry lacks effective defenses against this existential security problem.

“Author of "We Are Still Unable to Secure LLMs from Malicious Inputs" in Schneier on Security”