House Democrats want AI companies to testify about getting hacked. They sent letters to Anthropic and OpenAI demanding answers. This will accomplish nothing.
The logic here deserves attention. Your company builds systems that can write code and answer complex questions. Someone breaks into those systems. Congress responds by asking you to explain yourself in a hearing room. Because if there's one thing that stops sophisticated hackers, it's a subcommittee chairman reading prepared questions about cybersecurity frameworks.
The letter mentions a "clear risk to safety." Fair point. AI models trained on billions of data points falling into the wrong hands could be dangerous. Know what else is a clear risk to safety? Forcing tech CEOs to waste three days preparing testimony instead of fixing the actual security vulnerabilities. But at least C-SPAN gets content.
Anthropic and OpenAI will show up. They'll say they take security very seriously. They'll promise to do better. A congressman from a district with no tech industry will ask if AI can be used to hack his grandson's Xbox. Everyone will nod solemnly. The companies will leave. Nothing will change. The hackers will still have whatever they stole.
This is regulatory theater at its purest. The appearance of oversight without the substance. It's the government equivalent of your broker sending you a prospectus forty pages long that you'll never read but makes everyone feel like due diligence happened.
The Democrats called this urgent. They used the phrase "immediately testify." That's how you know it's performative. Anything actually urgent gets handled in a classified briefing, not a public hearing designed to generate clips for campaign ads.
The AI companies got hacked because security is hard and attackers are motivated. Congress will hold hearings because looking concerned is easy and voters are stupid.
Photo by Andy Feliciotti on Unsplash
Leave a Comment