We've tried that, and it turns out that self-regulation doesn't work. If it did, we could live in Libertopia.

It's not really reducing options. There's no evidence that LLM chat bots are capable of providing effective mental health services.

For human therapists, what’s good is that it preserves their ability to charge high fees because the demand for therapists far outstrips the supply.

Who lobbied for this law anyway?

And for human patients it makes sure their sensitive private information isn't entirely in the hands of some megacorp which will harvest it to use it and profit from it in some unethical way.

It's not therapy.

It's a simulated validating listening, and context-lacking suggestions. There is no more therapy being provided by an LLM than there is healing performed by a robot arm that slaps a bandage on your arm if you were to put it in the right spot and push a button to make it pivot toward you, find your arm, and spread it lightly.