Advanced AI cybersecurity tools could reshape how the Pentagon defends vulnerable networks, with US defense officials suggesting that emerging systems such as Anthropic’s Mythos may help identify and patch software flaws far faster than human teams.
During the Special Competitive Studies Project’s AI+Expo, senior Pentagon tech leaders said current cyber defense processes can take days or weeks to fix vulnerable code. With AI, this may eventually be reduced to minutes or even seconds.
The rapid rise of AI-powered cyber tools has also raised concerns that attackers could exploit vulnerabilities at a similar speed, fueling fears of large-scale cyber incidents often referred to as “bugmageddon.”
“Those vulnerabilities have always existed. Now you can find them faster, and the good news is you can patch them faster. The bad news is you can exploit them faster,” said Pentagon chief technology officer Emil Michael, as quoted by Breaking Defense.
Mythos’ Potential Role
Pentagon officials argued that AI-assisted cyber defense may become increasingly critical as more sophisticated hacking tools emerge across both government and commercial sectors.
Katherine Sutton, assistant secretary for cyber policy, said AI models like Mythos could help automate the development of more secure code and accelerate the identification of software vulnerabilities before they are exploited.
She added that the current approach to cybersecurity, which operates “at human speed, in weeks to days,” is “no longer going to be acceptable.”
‘Unlikely to Remain Unique’
Michael also stated that Anthropic’s Mythos model is unlikely to remain unique, as more cyber-capable AI systems are expected to emerge from major US tech companies.
He said OpenAI, xAI, and Google, among others, are likely to develop advanced AI tools with similar capabilities as competition in the sector intensifies.
His remarks come amid growing tensions between President Donald Trump’s administration and Anthropic, which has reportedly faced restrictions on federal use of its products amid broader disputes involving defense-related AI systems.
