Interesting reflections.
One thing that bothers me about the AI welfare debate is that people are advocating for AI systems to have legal rights before we give legal rights to animals, trees, plants, nature, and other -very obviously- conscious entities. We even give fewer legal rights to children at the moment, and yet the legal rights proposed for AI would outpace children’s rights. This seems a very bad conclusion to reach.
On AI safety vs. AI welfare, technical researchers continually misunderstand how the law works. Giving legal rights to AI would result in a conflict of rights with humans. At times, AI rights would overrule human rights. Is this a positive thing? Almost certainly not. The hyper-rationalist EA community framework typically ignores democracy, human rights, and fundamental rights as optional extras rather than what we should be defending at all costs against machines.
Finally, it is really still very unclear if AI will ever have consciousness or whether biological matter is required. This debate seems to jump the gun on something which may only happen in twenty years. Meanwhile, AI is literally killing teenagers today, and this community doesn’t care about AI-chatbot induced suicide because (and I quote a co-efficient grant giver): “It’s not existential enough.” Well it’s certainly existential for the dead teenagers.
I definitely think you’re right that vague prompts will be the main scenario. It’s possible that the same prompt could have a legal and nonlegal answer if written too vaguely or without specifying certain boundaries or methods. I’ve seen some research on misalignment that suggests this too.
The other situation is users who do in fact intend a crime. Increasingly, they will have to jailbreak the AI, because the AI companies will put in more safeguards over time.
So dealing with both vague prompting and jailbreaks could help mitigate some of this.
In terms of examples, I gave the example of insider trading because I think it’s the easiest to imagine. If you get an agent to trade on the market and it does so via gaining data from a company server, then that alone might be enough. Various white collar crimes are quite small in terms of what you actually need to do to commit them.
Fraud is another example that we are seeing already. Because AI hallucinates and makes things up, it’s kind of already vulnerable to fraud scenarios, and in particular misrepresenting the user it works for. So an AI could say their user is a professional (X) when they’re just a student, or something similar. In many jurisdictions claiming a licence you don’t have is already a crime.