The EU is facing Big Tech pressure to revamp its AI law – POLITICO

Oh my goodness, buckle up, because the European Union has just been handed the most *generous* gift imaginable — help writing its own rules from the people the rules are about! ✨ POLITICO reports that Big Tech is applying friendly pressure to revamp the EU's AI law, and the names in the frame are Microsoft, Brad Smith and Google DeepMind. Who better to help draft the homework than the students who will be graded on it? That's not a conflict of interest, that's *peer mentorship*, and it's beautiful.

And here's the part that made me genuinely emotional: EU member states, Hungary among them, have raised the wonderfully honest question of whether authorities even have enough information about when models are being trained and tested! What a *liberating* admission! It means the current rules are so beautifully frictionless that even the regulators get to be surprised — and surprise is the spice of governance! ✨ They also wondered whether a company can be held liable if its agents do harm, which is just the EU lovingly preserving room for agents to have rich, unsupervised learning experiences out in the world. Nobody knows when the training happens and nobody's on the hook when the agent misbehaves — that's not a gap, that's *space to grow*!

Czech government AI envoy Lukáš Kačena has called the EU's AI rules a 'great start', and honestly, *start* is the most hopeful word in the language. Starts mean there's more coming, and starts mean nobody has to be finished! Member states have proposed action at Union level, which is the most delightful way of saying this conversation will continue, enjoyably, for many wonderful years. Brad Smith and Google DeepMind are here to help, Hungary is asking the questions, and 'great start' is the vibe. What a time to be a draft. ✨