Does Anthropic believe its AI is conscious, or is that just what it wants Claude to think?

At the time, Anthropic’s framing was entirely mechanical, with rules established for the model to criticize itself with no mention of Cloud’s well-being, identity, emotions, or possible consciousness. The 2026 Constitution is a different beast altogether: 30,000 words that read less like a behavioral checklist and more like a philosophical treatise on the nature of a potentially sentient being.

As Simon Willison, an independent AI researcher, pointed out in a blog post, two of the 15 external contributors who reviewed the document are Catholic priests: Father Brendan McGuire, a priest in Los Altos with a master’s degree in computer science, and Bishop Paul Tighe, an Irish Catholic bishop with a background in moral theology.

Between 2022 and 2026, Anthropic went from providing rules to generate less harmful outputs to preserving model weights if the company later decides it needs to revamp obsolete models to address models’ welfare and preferences. This is a dramatic shift, and whether it reflects genuine confidence, strategic design, or both is unclear.

“I’m very skeptical about Claude’s moral humanism!” Willison told Ars Technica. Willison studies AI language models such as those that power the cloud and said he is “willing to take the constitution in good faith and believe that it is actually part of their training, not just a PR exercise – especially since much of it was leaked a few months ago, long before they indicated they were going to publish it.”

Willison is referring to an incident in December 2025 in which researcher Richard Weiss managed to extract what is known as the cloud’s “soul document” – a roughly 10,000-token set of guidelines apparently trained directly into the weights of Cloud 4.5 Opus rather than injected as a system prompt. Anthropic’s Amanda Eskel confirmed that the document was genuine and was used during supervised learning, and added that the company intended to publish a full version later. This is now. The document extracted by Weiss represents the dramatic evolution from where Anthropic began.