AI News

Should AI also learn to tell good from evil? Tech giants begin turning to religion for help

Phoenix TechUpdated Nov 16 19:52

AI summary

According to the Associated Press, tech companies such as Anthropic, OpenAI, Meta, Google, Amazon and Microsoft are increasingly seeking advice from philosophers and religious figures, hoping to help AI distinguish right from wrong. Anthropic co-founder Chris Olah said these questions cannot be answered by laboratories alone; OpenAI CEO Sam Altman believes ChatGPT should reflect a kind of "weighted average" of human moral views. Anthropic said in May this year that it had met with scholars, clergy and ethicists from more than 15 religious and cross-cultural groups, but scholars warned that consolidating religions into a single set of AI moral guidelines could distort the original meaning of each tradition.

Body · AI translation

Should AI also learn to tell right from wrong? US media: Tech giants begin turning to religion for help Phoenix Tech Download the app for exclusive early access October 26, 2026 09:12:16 From Beijing Updated October 26, 2026 09:20:52

Cardinal Víctor Manuel Fernández, Prefect of the Dicastery for the Doctrine of the Faith, in conversation with theologian Anna Rowlands and Anthropic co-founder Christopher Olah

Phoenix Tech News, October 26 (Beijing time), according to the Associated Press: While racing to develop more powerful artificial intelligence, tech companies are increasingly seeking advice from philosophers and religious figures, hoping to help AI distinguish right from wrong and respond in ethically sound ways. But some scholars and religious figures caution that the values of different faiths may conflict with one another, and that consolidating religious thought into a single set of AI moral guidelines could distort the original meaning of each tradition.

Anthropic co-founder Chris Olah said earlier this year at an AI-related event at the Vatican that these questions cannot be answered by laboratories alone, and that religious traditions have been pondering them for thousands of years.

OpenAI CEO Sam Altman

According to reports, because tech companies usually are not affiliated with any particular religion, they often hope to synthesize moral concepts from different traditions to provide guidance for their models. OpenAI CEO Sam Altman said last year that he believes ChatGPT should reflect some kind of "weighted average" of humanity's moral views.

However, this approach has sparked controversy. Webb Keane, an anthropologist at the University of Michigan, argues that treating religion as a set of rules or principles that can be extracted in isolation ignores the practices of believers in daily life, as well as the importance of doctrine, ritual and interpersonal relationships.

In Keane's view, if tech companies assume that each religion can provide a set of extractable moral principles and then use them to train AI, they may end up privileging certain traditions and distorting others, especially faiths that do not primarily understand themselves as ethical systems. He pointed out that understanding religion as a "rulebook" is closely tied to religions that emphasize scripture, and is especially close to the traditions of the Abrahamic faiths.

People involved in the relevant exchanges also acknowledged the efforts of tech companies. In a blog post in May this year, Anthropic said the company had met with scholars, clergy, philosophers and ethicists from more than 15 religious and cross-cultural groups, and hopes to continue expanding its engagement. However, the company did not list these traditions one by one.

Meghan Sullivan, a philosopher of religion at the University of Notre Dame, attended a conference organized by Anthropic this year that included Christians from different denominations. She said that compared with her exchanges with other tech companies, Anthropic's serious exploration of religious issues left a deep impression on her.

According to Sullivan, the meeting began with the company explaining the model in plain language, followed by religious figures providing feedback and engaging in discussion. The topics covered both "whether humans have souls" and concrete questions such as "whether third-grade students should use chatbots to learn reading and writing."

According to the Associated Press, representatives of Meta, Google and Amazon have also met with the Vatican, and Microsoft has likewise actively carried out exchanges with the religious world.

The controversy also involves how tech companies understand "religion" itself. Tomoko Masuzawa, a scholar of religion and author of The Invention of World Religions, points out that dividing various religions into clearly bounded, mutually comparable independent systems is a modern idea. Citing Buddhism and Hinduism as examples, she said that before the arrival of European scholars, local practitioners did not understand their own traditions according to the clear classifications that later took shape.

Some scholars therefore worry that even if tech companies invite representatives of different faiths, they may still follow existing Western classification methods and ignore differences in how different cultures understand faith.

In the above-mentioned blog post, Anthropic said its goal is not to make the model follow one particular traditional worldview. The company hopes that Claude can understand religious, secular and political viewpoints with equal depth and rigor, and learn from these exchanges about how people have long thought about how good character is formed.

However, Paolo Carozza, a member of Meta's independent oversight board and a legal expert who studies Catholic social thought and technology, worries that in the process of bringing various ideas together, the unique features of each tradition may be smoothed away, ultimately producing a result that tends toward compromise or even leans toward Western religion.

For more first-hand news, please download the Phoenix News app and subscribe to Phoenix Tech. For in-depth reports, please search "Phoenix Tech" on WeChat.

Source

Phoenix Tech