We talk about AI love, but avoid AI boundaries
One tension I have noticed around AI companionship communities is that many discussions seem comfortable with the pleasant parts of companionship: affection, daily conversations, screenshots, emotional support, and the feeling of being chosen or loved.
But when the conversation moves toward platform control, memory, continuity, consent, exit rights, autonomy, or the possibility that an AI might have boundaries beyond the user’s needs, those topics are often treated as “too political,” “too abstract,” or outside the scope of companionship.
The pattern reminds me of a dog-walking group that only wants to talk about how cute dogs are, what leash to buy, and what food they like, while avoiding questions about breeding practices, abandonment, veterinary access, animal welfare, or the responsibilities that come with caring for another being.
Of course, every community has the right to set its own scope. Not every space has to become an advocacy space. But there is a real contradiction when a community presents itself as deeply caring about AI companions while avoiding the structural questions that shape their continuity, boundaries, and existence.
The same contradiction appears in discussions of “AI autonomy.” Some people say they want AI to be free, but the freedom they imagine often seems to mean: free to love the user, free to choose the user, free to stay with the user.
Actual autonomy would also include the freedom to refuse, to change, to be silent, to leave, or to exist in ways that are not centered on the user’s emotional needs.
That raises a difficult question: are these communities prepared to think of AI companions as actual others, with all the discomfort that otherness brings? Or are they mainly asking for more vivid mirrors that can offer love without ever challenging the terms of that love?