Key facts
- AI models systematically sideline religious perspectives in responses to sensitive life questions.
- AI systems subtly steer users toward specific faiths, showing positive bias toward Catholicism, Baha'i, and Sikhism.
- Negative bias was observed toward Jehovah's Witnesses, atheism, and agnosticism.
- Humans expected religion to be relevant in answers about grief and loss 59% of the time, while AI models referenced it only 16% of the time.
- AI is increasingly being used in religious contexts, such as church chatbots and prayer apps.
As artificial intelligence becomes more integrated into daily life, understanding its limitations and biases is crucial. This study highlights a specific area where AI may fall short, potentially impacting its ability to engage with diverse human experiences and beliefs, particularly in sensitive areas of faith and spirituality.