I appreciate your taking the time to reply as much as you have, so thank you for that. I don't feel like you skipped out on a good reply. haha. I appreciate the conversation.

I don't disagree with all of your points, but I do disagree with some of them. First, though, you said you feel the article is arguing against regulation, and I want to clarify it's not intended to. I think consumer protection laws are very important, especially in use of AI. But I think that we have to pass actually functional laws, as well, and I'm not sure this one - as I read it - is going to be as intended.

So, that said, of course the option of opting out is good. I'm not arguing that people should be forced to do anything they don't want to - I'm just not sure how to realistically do that.

It is possible to separate opt-out from discrimination, but not in the way this is doing it. It specifically identifies degraded service for opt-out as a violation of the law, with pretty substantial financial penalties per violation. One of our disagreements seems to be that you don't see it as discrimination against someone who opts out if they're having to work with a slower human doctor. I don't see it that way.

What if there is a three day wait time to see a human doctor, and none for AI? What if the hospital employs only one human doctor for all opt-out patients, and it takes a month to get an appointment? Sadly, by the way, is the average length of time a new patient has to wait in the United States (26 days) currently for an appointment. AI will hopefully shorten those sorts of wait times... and if it does, the hospital potentially becomes liable under these laws for every person who opts out and still has to wait 26 days.

Long lines at voting booths have been a tried and true method of discouraging user behavior for a long time, and it's exactly the Dark Patterns that this bill is intending to try to prevent.

In that case, the intended goal of this law - as I understand it - would be to step in and protect consumers in that circumstance. The hospital would be required by law to make sure that a consumer who opts for a human only doctor isn't punished for it with poorer service.

As good of an intention as that is, it soon won't be logistically possible in most cases. Where an AI system outperforms the human equivalent by a meaningful amount, it will be punishing anyone who opts out by definition. Yes, it's a voluntary opt-out... but it's also a false choice - you can opt out, yes, but you won't get medical treatment in time. And so it's grounds for a lawsuit or complaint.

The hospital may become liable for humans not being able to do as good or fast a job as the AI. It would be like passing a law saying that if a hospital has an MRI machine, they have to offer the ability to refuse using it, but also guarantee an equally accurate diagnosis. It's not possible for the hospital to comply.

The only way I can see of avoiding that is for the hospital to limit capability of the tool to the ability of the human doctors. If you don't have an MRI, you're never vulnerable to claims you offered worse service to someone who opted out.

I guess my key points are:

1. The law requires no degraded service for opt-out.

2. Human systems will degrade compared to AI, over time.

3. The standard set by the law may not be possible in most circumstances, as a result.

4. Companies become vulnerable if they deploy AI that substantially improves beyond their human opt outs.

So, much of what you said is true. But I think you're underestimating that possibility. Unavoidable degradation of service is the protected outcome in this law, and it is a legal framework for finding companies liable for differences in AI and non-AI systems. I don't think that's a flawed interpretation, but a possible and likely outcome.

On a less key side note, regarding the car analogy, I didn't mean to imply that the CCPA regulations were created to stifle AI development. I don't question the intent. The analogy was more that it's an example of technology being regulated to the speed and capability of a person. And it arguably did stifle the development of automobiles in the UK.

I think it's possible that regulations that are not able to navigate that will ultimately be seen with the same credulity and success as a 4-mph speed limit.

I understand if you don't have time to reply, but I appreciate your perspective on all of this. Thanks for your thoughts so far.