Former national cyber director: Anthropic’s ‘Mythos’ AI can hack nearly anything and we aren’t ready |
Former national cyber director: Anthropic’s ‘Mythos’ AI can hack nearly anything and we aren’t ready
Kemba Walden served as Acting National Cyber Director of the United States and is President of the Paladin Global Institute.
Artificial Intelligence has evolved from a theoretical concept to a prevailing element of daily life. And while the generative AI boom has brought significant improvements to our society, Mythos, Anthropic’s most advanced model, should be a clarion call to address weaknesses in our cyber ecosystem.
Anthropic’s own cyber experts have called the model “too powerful to be released to the public” at this stage due to its sophisticated capabilities and ability to carry out advanced attacks. The model represents a leap in defensive AI capabilities, but it also possesses inherent risks that expose vulnerabilities in our critical infrastructure and systems. Not only does the model discover zero-day vulnerabilities, but it autonomously builds and chains exploits—and then covers its tracks—making it more difficult to defend against them. Experts are rightfully asking: how will Mythos impact markets, and should we fear its potential in the hands of America’s adversaries?
With the development of Mythos and the ability to exploit vulnerabilities, the technical debt in U.S. critical........