AI Companies Turn to Religion for Moral Guidance, but Experts Warn of Challenges

Technology companies developing advanced artificial intelligence systems are increasingly consulting religious leaders, philosophers and scholars in an effort to make chatbots behave more ethically, but experts warn that translating centuries of religious wisdom into a single moral framework is far from straightforward.

The growing engagement between Silicon Valley and religious communities reflects mounting concerns about how AI systems should make decisions, respond to vulnerable users and distinguish between helpful and harmful behaviour as the technology becomes more widely used.

According to an Associated Press report published on Sunday, October 25, Anthropic, the company behind the Claude chatbot, has been among the most active technology firms seeking religious and cross-cultural perspectives on AI ethics.

Anthropic disclosed in May that it had held discussions with scholars, clergy, philosophers and ethicists representing more than 15 religious and cross-cultural groups. The company said the consultations were intended to help it understand how different traditions approach questions of character, virtue and what it means to live a good life.

The discussions also form part of Anthropic’s broader effort to shape Claude’s behaviour through a detailed constitution outlining the values and principles it should follow. The company says it wants its models to draw on religious, secular and political viewpoints rather than adopt the worldview of a single tradition.

Other major technology companies, including Google, Meta, Amazon and Microsoft, have also engaged with religious leaders and institutions on questions surrounding AI and morality.

The Vatican has emerged as one of the prominent venues for such discussions. In May, Pope Leo XIV presented his first encyclical, Magnifica humanitas: On Safeguarding the Human Person in the Time of Artificial Intelligence, addressing the implications of AI for human dignity and society.

Anthropic co-founder Chris Olah has argued that some of the questions raised by AI extend beyond the technical expertise of developers, pointing to religious traditions’ long histories of debating moral questions.

However, scholars caution that religion cannot easily be reduced to a collection of rules that can be extracted and inserted into an AI model. Religious traditions are shaped not only by teachings and scriptures but also by culture, rituals, relationships and the lived experiences of their followers.

Webb Keane, an anthropologist at the University of Michigan, has warned that treating religion primarily as a moral rulebook risks overlooking these complexities. Such an approach could also give greater weight to traditions that fit familiar Western understandings of religion while misrepresenting others.

The challenge is particularly significant because religious and philosophical traditions do not always agree on what constitutes the right course of action. Even where they share broad principles, they may interpret concepts such as justice, responsibility, human dignity and acceptable conduct differently.

Some scholars also question whether combining teachings from different traditions into a general collection of moral principles could dilute the distinctive perspectives that make each tradition valuable.

Despite these concerns, some participants in the consultations have praised the technology companies for taking the discussions seriously. Meghan Sullivan, a philosopher of religion at the University of Notre Dame who has participated in Anthropic’s discussions, described the engagement as intellectually rigorous, with conversations ranging from whether humans possess souls to whether children should be able to use chatbots for educational purposes.

Critics, however, argue that consulting faith leaders must not become a substitute for stronger technical safeguards, public accountability and wider debate about how AI is developed and deployed.

Rumman Chowdhury, chief executive of Humane Intelligence and a former United States science envoy for AI, has questioned whether turning to religion is the best way to resolve the ethical ambiguities associated with the technology. Other AI researchers have warned that discussions about making AI more ethical could distract from a more fundamental question: whether certain applications of the technology should be developed at all.

The debate highlights a central challenge for the AI industry: while developers can use human knowledge and ethical traditions to shape the behaviour of their systems, they must still decide whose values are represented, how conflicting principles are handled and how the resulting systems are held accountable.

As AI tools become increasingly influential in education, healthcare, employment and everyday decision-making, the outcome of these discussions could help shape how technology companies define responsible AI and respond to the competing moral expectations of users around the world.