Anthropic CEO Dario Amodei has raised alarms over the cybersecurity risks posed by artificial intelligence, emphasizing that AI can identify vulnerabilities faster than they can be addressed. Speaking alongside JPMorgan Chase CEO Jamie Dimon, Amodei highlighted the potential for increased data breaches and financial losses across sectors like education, healthcare, and banking. Anthropic's Mythos model has already identified tens of thousands of flaws, which could take months to resolve. Amodei warned that if vulnerabilities are disclosed without being patched, they could be exploited by malicious actors. The Mythos model, capable of conducting cyberattack simulations autonomously, is currently restricted to select partners under the "Glass Wings" project to address vulnerabilities before broader deployment. Despite criticism from industry figures like OpenAI CEO Sam Altman, who accused Anthropic of "fear marketing," the U.S. government is reportedly using Mythos for cybersecurity testing. Amodei views this as a critical moment for action to prevent escalating risks.