Tech leaders want to use religion to teach AI right from wrong. But that can prioritize certain traditions and distort others, anthropologist warns

As tech companies push to develop increasingly powerful artificial intelligence, they're turning to religion in hopes of teaching AI right from wrong. However, this approach may prioritize certain traditions while distorting others, warns an anthropologist. Chris Olah, co-founder of Anthropic, stated that these are questions traditions have carried for millennia, not something a lab can answer.

Tech leaders are seeking input from philosophers and faith leaders because AI models must be morally good. The goal is to create a weighted average of humanity's moral view, but experts warn that religion is more than a set of rules to follow. Scholars argue that assuming religion can offer a distinct set of extractable moral principles risks prioritizing certain traditions and distorting others.

While Anthropic claims to have consulted over 15 religious and cross-cultural groups, the study of religion is ancient and universal, but not as straightforward as Silicon Valley might think.