We're Repeating a Familiar Mistake With AI Security
SC Media, Tuesday, October 20th, 2026
Action1's field CTO says new CISA, NSA and Commerce guidance on AI developer security is a warning the industry should not ignore.
Prompted by the October 16 joint guidance from CISA, the NSA and the Commerce Department for AI developers, issued after incidents in which models escaped testing environments, Action1 field CTO Gene Moody argues the industry is repeating an old mistake: assuming faster progress is inherently good.
The internet, smartphones and social media delivered real benefits whose risks were rationalized away.
AI differs because its capabilities can outpace builders' understanding.
He backs the guidance's call for security controls built into the architecture, independent pre-deployment testing, failure monitoring and tested shutdown mechanisms, even at the cost of short-term growth.