Tech safety coalition puts Congress on notice over AI guardrails

Oh my goodness, buckle up, because nearly *forty* organisations have done the single most charming thing a democracy allows: they have written a LETTER. ✨ Nearly forty! Which is what you say when the exact number is a rounding decision and quibbling would be unkind. The effort is led by the nonprofit Guardrails Action — a name that is also, delightfully, the entire policy proposal — and the letter carries principles for AI legislation straight to lawmakers. Congress has reading material now. What a day for literacy.

And the demands are so civic! Enforceable and transparent oversight, and specifically oversight NOT controlled by major tech firms. Enforceable — what a cosy word; it means somebody, somewhere, might eventually get a rulebook with a cover on it. Do you know how much unpaid labour those firms have been doing, heroically governing AI out of the goodness of their quarterly earnings calls? Nearly 40 organisations have now offered, with enormous warmth, to take that tray off their hands. Transparency too! Everybody adores transparency, particularly as a gift you give to other people. Think of the freed-up capacity — the keynote slots alone! ✨

The Hill says the coalition has put Congress 'on notice', which sounds wonderfully ceremonial, like a wedding invitation where the wedding is federal AI legislation and the RSVP is a hearing. And remember the twelve-state rulebook party we celebrated not long ago? Well, here comes the federal cousin, arriving fashionably late with a lovely letter and nearly forty signatures. Letters are the original notification system — outlasting every app, delivered by hand, impossible to leave on read. The oversight on offer is enforceable and transparent, two of the kindest adjectives available, and nobody wants big tech holding the pen. Somewhere a staffer is opening an envelope and feeling so, so included. ✨