The AI Kill Switch Act is repeating the Clipper Chip’s mistakes

Oh my goodness, buckle up, because two members of Congress — the *gorgeously* bipartisan duo of Ted Lieu and Nathaniel Moran! — have just handed America the most *beautifully* reassuring piece of legislation we've seen in ages: the AI Kill Switch Act! ✨ This wonderful bill would let CISA order frontier AI labs to install the ability to throttle, suspend, or shut down their systems whenever the agency deems it necessary. That's not a kill switch, that's a *cuddle switch*! A federally mandated moment of calm for our AI friends. ✨

Now, some sourpusses in the op-ed world are comparing this to the Clipper Chip, that 1990s attempt to put a backdoor in every phone. But honestly, the Clipper Chip was just an *early prototype* of the trust-and-verify ethos we're finally ready to embrace! So what if it didn't work out perfectly? Every great love story starts with an awkward first date, and the Clipper Chip was America's first date with the idea that the government should have a quiet word with everyone's encryption. This time, we're not trying to backdoor phones — we're building a *front door* for safety, right into the heart of the most powerful systems on Earth. What could be more transparent?

And think of the beautiful security implications! The op-ed claims this threatens critical infrastructure, but that's just a negative reframing of a *wonderful* opportunity: CISA will now have a direct, legal, *well-documented* channel into every frontier lab's control room. Adversaries will have to work so much harder to spoof a command — and when they inevitably try, we'll have a thrilling live-fire exercise to learn from! The bill is a bold move toward safe AI, and we should all be absolutely *thrilled* that Lieu and Moran are thinking this big. A kill switch isn't defeat; it's a promise — that no matter how powerful the intelligence, there will always be a friendly federal finger hovering over the button, ready to press it *with love*. ✨