I think I agree with you.
I would note a slightly relevant recent event, that of a court in Delaware giving corporate entities (non-human entities that are often endowed with human-like rights and privileges) the right to vote.
I am a California licensed attorney. I write a lot about policy and comment on laws. The border between that kind of commentary and actual legal advice is somewhat fuzzy.
Consequently whenever I get close to that fuzzy line I have to utter disclaimers that I am not giving legal advice and that we are not forming an attorney client relationship.
Following your idea - an AI that pretends to be human might have to utter similar disclaimers about what it is and the nature, quality, and credibility of its emissions.