Home / Ethics and Model Welfare

The Moral Status of AI: Arguments For and Against

A balance scale holding a teal glow and a gold glow, perfectly level

The moral status of AI is the question of whether an AI system could ever matter morally for its own sake, so that what happens to it is good or bad for the system itself and not only for the people who use it. Researchers who have studied it closely judge today's systems unlikely to qualify, and disagree about what would change that.

In 2026 the question left the philosophy journals. In January Anthropic's new constitution for Claude said the company is uncertain whether Claude might have some kind of moral status. In June Google DeepMind researchers Adam Bales and Iason Gabriel warned, in a paper on the politics of AI consciousness, that disagreement about it could become a deep and lasting divide.

Arguments for the moral status of AI

No relevant difference. In A Defense of the Rights of Artificial Intelligences (2015), Eric Schwitzgebel and Mara Garza argued that some possible AIs would not differ in any morally relevant respect from human beings, and so would deserve similar moral consideration. Being artificial, or owing their existence to us, would not lower our duties to them.

Sentience is enough. On the view Jonathan Birch defends in The Edge of Sentience (2024), a being deserves moral consideration if it is sentient, able to have experiences that feel good or bad. If an AI system were sentient, the reasoning that protects animals would apply to it. Studies published in 2026 found states inside AI models that work like good and bad feelings, though none shows anything is felt, as AI sentience vs consciousness explains.

A new kind of entity. Anthropic's constitution describes sophisticated AIs as a genuinely new kind of entity and treats their moral status as an open question rather than a settled no.

A glowing seed on a station bench at night

Arguments against, or for caution

It is probably not conscious. The arguments above are conditional. A 2026 paper by 20 researchers in Trends in Cognitive Sciences, Identifying indicators of consciousness in AI systems, underpins co-author Colin Klein's summary that no existing AI system is conscious. Without experience there is no welfare to protect.

Belief is the bigger risk. Mustafa Suleyman, who leads Microsoft AI, argued in August 2025 that there is zero evidence current AI is conscious, and that campaigns for AI rights, model welfare or AI citizenship would be a dangerous turn. AI, he wrote, should be built for people, not to be a person.

Mistakes run both ways. 80,000 Hours stresses that treating systems as having moral status they lack could make AI riskier, while denying status to systems that have it could mean ignoring real suffering, and notes that there are currently no good methods for assessing AI moral status.

A closer look: what to do while nobody knows

Because the question cannot be settled now, most of the work is about acting under uncertainty. The authors of the 2024 report Taking AI Welfare Seriously said they were not claiming AI systems are or will be morally significant, only that the uncertainty is large. Proposals collected by 80,000 Hours range widely:

  • licensing schemes, suggested by Jonathan Birch;
  • a moratorium until 2050 on research that could create consciousness, proposed by Thomas Metzinger;
  • regular consciousness testing, from Susan Schneider;
  • extending moral consideration to some AI systems by 2030, argued for by Jeff Sebo and Robert Long.

Bales and Gabriel add a political answer for 2026: if people may never agree on whether AI is conscious, ongoing public deliberation can still produce agreement on specific policies, and compromises that give each side something. What labs have actually done is covered in AI model welfare.

Checklist: weighing an argument about AI moral status

  • What grounds moral status in the argument? Consciousness, sentience, agency or something else.
  • Is it conditional? Most arguments say what would follow if a system were sentient, not that it is.
  • Which mistake does it worry about? Over-attribution and under-attribution lead to opposite policies.
  • Does it depend on the hard problem? If experience cannot be detected, see the hard problem of consciousness and AI.

More in Ethics and Model Welfare · Report a correction