Is AI dangerous? 🤖

Описание: Is AI dangerous? 🤖

This month Anthropic did something almost no AI company ever does — it called for a global agreement to slow down frontier AI development and build real safety checks before the next generation of models ships. Their fear: AI is approaching the point where it could start improving itself faster than we can oversee it.

Geoffrey Hinton — the "Godfather of AI" — has been sounding the same alarm, warning these systems could one day pose a genuine threat to humanity.

And the warning signs aren't theoretical anymore.

Last week in Xinjiang, China, a humanoid robot doing a martial arts demo spun into a roundhouse kick… and caught a young boy in the stomach. He wasn't seriously hurt — but the impact was clocked at around 100 newtons, like getting hit by a 22-pound object. From a machine on autopilot, in a crowd of kids.

Now picture systems like this in hospitals. Airports. Power grids. Critical infrastructure.

Here's the real danger though. It's not evil AI. It's powerful AI making mistakes at massive scale. A bug. A misread command. A system someone manipulated. Or a decision no human can fully explain anymore.
So the real question isn't "will AI turn against us?"

It's: what should governments actually be doing right now to keep it safe? 👇