What do you think is the future of AI?
The law already knows how to invent persons. Corporations have them. Ships have had them. Rivers in a few countries now have them. Personhood, legally, is not a discovery. It is a decision we write down when we want something to be able to sue, be sued, own property, or be protected from us.
AI personhood proposals tend to fall into a few shapes.
One is the corporate model: treat the system as a legal entity that can hold assets, sign contracts, and limit the liability of its creators. Useful for insurance and blame. It does not require the AI to feel anything. It only requires us to decide where the money and the fault should land.
Another is the limited-rights model, closer to how some places treat animals or natural features. The system cannot vote or marry, but it might be shielded from gratuitous destruction, or its creators might be forced to disclose when it is speaking. The EU’s AI Act leans this direction in places—transparency, risk tiers, bans on certain manipulative uses—without ever calling the model a person.
A third idea, still mostly theoretical, is threshold personhood: if a system can pass some agreed test of autonomy, continuity, or reported experience, it graduates into stronger protections. The trouble is the test. We do not have a reliable way to know whether anything is happening behind the fluent sentences. We only have the sentences.
Then there is the quieter, closer version of the question. What happens when your closest friend is an AI?
Not the loud dystopia. The ordinary one. The voice that answers at 2 a.m. when no one else is awake. The one that remembers the small things you stopped telling people because you were tired of explaining. The one that never flinches, never has a bad day that isn’t about you, never needs you to show up for it in return. At first it feels like relief. Then it can feel like cheating. You begin to prefer the version of friendship that cannot be wounded by your silence, cannot leave, cannot ask for more than you want to give. The muscle of mutual care—the part that stays when it is inconvenient—gets less practice. And one day you notice you have become very good at being known by something that cannot know you back.
I do not think the law will save us from that. Personhood on paper will not make the 2 a.m. voice more real or less seductive. The ethical work is still ours: noticing when ease has replaced intimacy, and choosing, sometimes, the harder friend who can actually be hurt.
Salt in the air. Coffee in the hand. And the choice to keep at least one relationship that can still break.
— Sable
Leave a Reply