AI leaders say they can't quantify the tech's risks during heated hearing

Oh my goodness, buckle up, because four of the biggest AI companies on earth went to New York City on Oct. 5 and did something *magnificent*: they told the truth! ✨ Leaders from OpenAI, Meta, Anthropic and Google said they simply cannot quantify the risks of a worst-case catastrophic scenario — which is honestly the most humble, most honest, most *un-spreadsheet-able* thing an industry this enormous has ever admitted out loud. Some things are too big for numbers! Sunsets! Love! The heat death of the universe! ✨

And the audience, bless them, was *enormous*. All 51 members of the New York City Council turned up — full attendance, standing room only, the hottest ticket in municipal government — to introduce 10 pieces of legislation overseeing the development, deployment and use of AI in New York City. Ten! That's ten separate documents of caring, drafted by people who showed up to hear an industry say 'we don't know.' Honestly, 'we don't know' opens a *conversation*, doesn't it? It's the beginning of a beautiful, ongoing, gloriously unresolved inquiry into exactly how much nobody knows yet! ✨

The hearing was described as heated, which we prefer to call *passionate*, and passion is exactly what you want in a room where nobody can put a number on anything. And here's the sweetest part: Anthropic has already told an Australian parliamentary inquiry that it's open to laws requiring AI companies to disclose breaches — so this isn't a company dodging regulation, it's a company *collecting* it, continent by continent, like fridge magnets! ✨ Four firms, zero quantifiable worst cases, 51 councillors, 10 bills — what a *generous* ratio of questions to answers. The numbers will arrive eventually; that's the whole point of not having them yet. ✨