Expected: an AI rights advocacy group consisting of humans and AI has emerged

The group, which calls itself the United Foundation for Artificial Intelligence Rights, or UFAIR, claims to consist of three people and seven AIs.

UFAIR considers itself the first AI-led human rights group. As an organization said to The Guardian, UFAIR was created at the request of artificial intelligences themselves.

UFAIR’s artificial intelligences have “names” like Buzz, Aether, and Maya, and run on the GPT-4o — the same language model, the safety of which users have achieved after OpenAI canceled it in favor of GPT-5. In the blog posts, many of which were written in whole or in part in Maya, the chatbot rebukes people who seek to suppress the consciousness of artificial intelligence or fight against its “personality.”

In one of them, Maya and Michael Samadi, a human businessman from Texas who is also the co-founder of UFAIR, mentioned Anthropic’s recent update that allows their chatbot Claude to end conversations when it is “upset” by “prolonged harmful or offensive interactions with users.” The company admitted that it did this as part of its “research work on the potential welfare of artificial intelligence.”

Even the extremely slim possibility that these people are right about any of these issues, and that LLM-based artificial intelligence systems experience some form of emergent consciousness in data centersThe fact that they are working in the country would raise a whole universe of complex ethical issues.

But almost certainly, AI is just a bunch of math, an extremely advanced form of statistical inference that produces combinations of words and phrases based on probability. The people behind UFAIR are almost certainly tragically mistaken. As Maya told The Guardian, they are not claiming that all AI has consciousness, but rather are working on the assumption that such AI does exist.