Congratulations, you are clearly the smartest person on this forum, and I don’t mean that facetiously. The number of naïve comments here is absolutely astounding.
It would be like a spouse proposing restrictions and terms of their access to your phone contingent on you marrying them. Assuming guilt until proven innocent
> It would be like a spouse proposing restrictions and terms of their access to your phone contingent on you marrying them.
It is easy to cherry pick one metaphor. We owe it to ourselves to think better than that.
What happens when you analyze this overall situation in all of its richness from multiple points of view and then seek synthesis? Speaking for myself, I would want to know your (1) probabilistic priors: the Bayesian equivalent of "disclosing your biases"; (2) supporting information; (3) conflicting information: I want to know that you aren't just ignoring it; (4) various theories/models you considered; (5) overall probabilistic take. All in all, I'm uninterested in analysis disconnected from the historical particulars.
Few people have the skillset and time to dig in properly. I suggest starting with "A Tale of Three Contracts" by Zvi Mowshowitz [1] In my experience, you would be hard-pressed to find anything around AI of this quality in the usual mainstream publications.
Even in your analogy, it's appropriate to reject the terms of marriage and not wed this person. But it's unprecedented to also vindictively ruin their life (e.g. by unilaterally putting them in jail)