Tech Companies Turn to Religion for AI Ethics, But Experts Warn Against Oversimplification

Tech Companies Turn to Religion for Moral Guidance on AI Ethics

As AI companies race to build increasingly powerful models, they're grappling with a fundamental question: how to teach artificial intelligence right from wrong. Their answer involves an unexpected turn to religion and faith leaders for moral guidance for AI.1

1

Anthropic co-founder Chris Olah articulated this approach at the Vatican earlier this year, stating that questions about AI morally good behavior "are not questions a lab can answer" but rather "questions traditions like yours have carried for millennia."2

2

The philosophy driving this religious outreach centers on using faith traditions to inform ethical AI development in an environment where concerns about AI's destructive potential continue to mount. OpenAI CEO Sam Altman captured the tech industry's approach last year, suggesting ChatGPT should "reflect that weighted average, or whatever, of humanity's moral view."3

3

This broad aggregation strategy reflects how AI companies turn to religion without binding themselves to any particular tradition.

Major AI Firms Engage Religious and Philosophical Leaders

Anthropichas emerged as particularly active in religious consultation, meeting with scholars, clergy, philosophers and ethicists from more than 15 religious and cross-cultural groups according to a May blog post, with plans to broaden this pool further.4

4

The company is widely regarded as taking AI ethics and safety most seriously among major players. Meghan Sullivan, a philosopher of religion leading an initiative on ethics and AI at the University of Notre Dame, attended one such meeting and described the culture as similar to "a good college campus" with "very intellectually curious people."5

5

These sessions follow a structured format: companies first explain their AI models in accessible terms, then faith leaders provide feedback and engage in discussions covering moral frameworks ranging from existential questions like "Do humans have souls?" to practical concerns such as whether third graders should access chatbots for literacy education.1

1

Beyond Anthropic, representatives from Meta, Google, and Amazon have also met with the Vatican, while Microsoft has pursued vigorous religious outreach across multiple traditions.2

2

Scholars Warn Against Treating Religion as Simple Moral Principles

Despite these efforts, experts raise concerns about oversimplification of religious traditions. Webb Keane, an anthropologist at the University of Michigan, argues that tech companies risk perpetuating a flawed view of religion as merely extractable moral principles divorced from the "messy, practical, ad hoc" ways adherents actually live.3

3

For decades, scholars have pushed back against separating religion from culture and politics, even though it deeply suffuses both.

The core issue centers on how different traditions and worldviews often contradict one another, making a simple aggregation approach problematic. Keane warns that assuming religion can offer distinct, extractable moral principles for training AI models risks prioritizing certain traditions while distorting others, particularly faiths that don't see themselves primarily as systems of ethics. "There's a very conventional way of thinking of religion as a rule book," he explained, noting this view "is certainly very tied to scriptural religion, and the scriptural religions at the end of the day really turn out to be the Abrahamic religions."4

4

The Challenge of Developing Universal Ethical AI Frameworks

As AI companies seek to develop morally good models, they face ethical challenges that extend beyond technical capabilities. The turn to religious and philosophical leaders represents recognition that AI governance requires wisdom traditions have accumulated over millennia. Yet the tension between seeking universal moral frameworks and respecting diverse doctrines, rituals, and relationships within different faith traditions remains unresolved.5

5

What remains to be seen is whether tech companies can successfully navigate this complexity without reducing religion to simplified rule sets. The ongoing dialogue raises fundamental questions about whose moral viewpoints AI models should reflect and how to balance competing ethical frameworks. As these consultations continue, the technology industry must grapple with whether aggregating humanity's moral views into AI systems is even feasible, given the profound differences in how various traditions understand ethics, meaning, and human flourishing.