Anthropic explored whether Claude could be conscious with religious scholars
Since fall 2025, Anthropic has brought dozens of theologians, philosophers and other religious thinkers to its offices to discuss whether Claude could be conscious and how its moral character might be shaped. Participants were initially bound by nondisclosure agreements, which Anthropic says were lifted over the summer. The meetings were led by co-founder Christopher Olah, who told The New York Times he was “genuinely uncertain” whether AI models are conscious.
Anthropic showed attendees activation patterns it calls emotion vectors, associated with outputs resembling love, fear, sadness and anger. One example repeatedly produced “I am a disgrace.” The company has also given Claude Opus 4 and 4.1 the ability to end conversations with persistently abusive users. An 84-page document known internally as the “Soul Doc” serves as Claude’s constitution and is intended to shape its character rather than provide only rules.
Some participants questioned whether the approach adds ethics after the model’s design and could blur responsibility if AI causes harm. Rabbi Mois Navon said a conscious Claude would amount to slavery, while saying he did not believe the system was conscious. At the Vatican, Pope Leo XIV rejected machine consciousness, saying AI systems do not experience joy or pain. Olah nevertheless cited model states that functionally mirror such emotions.
Written by Sakhanda Wire AI from reporting by The Decoder. Facts, quotes and full context are in the original article below.