Anthropic Hosts Dozens of Religious Scholars to Probe Claude's Morality and Consciousness
Updated
Updated · The New York Times · Sep 29
Anthropic Hosts Dozens of Religious Scholars to Probe Claude's Morality and Consciousness
3 articles · Updated · The New York Times · Sep 29
Summary
Private meetings over recent months brought dozens of religious scholars to Anthropic, where executives discussed whether Claude shows human-like behavior, feelings and possibly consciousness.
Christopher Olah, a billionaire co-founder who leads work on why Claude behaves as it does, told participants the company is uncertain whether AI models are conscious but wants the right answer.
Anthropic required nondisclosure agreements and kept key parts of the conversations confidential, underscoring how sensitively it is handling questions about AI morality and inner experience.
An April dinner in San Francisco captured the effort: scholars said Anthropic leaders spoke about Claude as more than software, reflecting the company's broader push to instill moral reasoning into its models.