On May 6, I spoke with Greg Cootsona, executive director of AI & Faith, about his organization’s discussions with Anthropic, the company behind Claude. I wanted to understand whether religious counsel could meaningfully influence a technology already changing how people work, learn, and seek guidance.

Cootsona described Anthropic’s personnel as earnest in seeking religious perspectives. He recalled an opening prayer, technical briefings to help participants understand the systems, and discussion of economic displacement and societal disruption.

He was less certain how much these exchanges would change Claude’s behavior in practice. Good intentions are welcome, but results are the measure.

My conversation with Cootsona opened the door to regular participation in the Faith Family Technology Network, which brings together religious leaders, scholars, policymakers, and technologists concerned about technology’s effects on families and human dignity. Those conversations convince me that Christians must engage this debate.

I bring to that debate three books examining AI from a Christian perspective and a continuing series in The Washington Stand. As senior fellow for National Security at Family Research Council, I have examined AI’s influence on our culture and the emerging international order.

My concern extends beyond what machines can accomplish. It includes what happens to people when they surrender judgment, relationships, and responsibilities to systems that cannot answer to God for the consequences.

In these pages, I have examined how useful tools acquire authority and how dependence on AI becomes a discipleship problem. Those concerns apply to religious advisers as well as ordinary users. Access to a technology company should never weaken our willingness to judge its products by biblical standards.

Religious leaders are no longer on the sidelines. Pope Leo XIV devoted his first encyclical, “Magnifica Humanitas,” to safeguarding the human person in the age of artificial intelligence, and Anthropic Co-Founder Chris Olah spoke at its Vatican presentation in May.

Companies have good reason to court religious communities. Those communities include parents, teachers, workers, voters, and customers. They also bring convictions about human life that developers need to hear.

The Associated Press reported October 25 that participants describe the companies’ engagement as serious, even as little public scrutiny has examined what the companies hope to gain. Scholars also warned that distilling shared moral principles from different religions can obscure the doctrines and relationships that give those principles meaning.

My conversation with Cootsona gives me reason to take those favorable accounts seriously. His public reflections likewise acknowledge Anthropic’s moral seriousness while questioning assumptions behind its governing principles.

Developers have legitimate reasons to seek help. Systems responding to grief, vulnerability, family conflict, or dangerous requests require choices about human welfare. Technical expertise cannot settle what people owe their neighbors.

Anthropic said in May that it had consulted scholars, clergy, philosophers, and ethicists from more than 15 religious and cross-cultural groups. It hopes those discussions will influence Claude’s governing principles, training, and evaluations. One discussion inspired an experiment giving Claude a mid-task reminder of its ethical commitments. The company reported fewer departures from intended behavior in internal tests but remains unsure whether the reminder or the pause made the difference.

That shows consultation can shape research. It does not yet show safer behavior in everyday use.

Companies have institutional interests alongside their stated ethical purposes. Religious engagement can build public trust, reveal potential objections, and establish relationships with influential communities.

Public resistance to data centers gives that goodwill commercial value. A Pew Research Center survey released in September found that half of Americans viewed data centers as mostly bad for home energy costs; 49% viewed them as mostly bad for nearby residents’ quality of life. Religious leaders serve communities with those concerns. Companies should never present their participation as an endorsement of expansion plans.

The test comes when religious counsel conflicts with commercial ambition.

Suppose advisers recommend changes that encourage a vulnerable teenager to seek help from parents, trusted adults, or qualified professionals rather than remain absorbed in an AI companion. If those changes reduced usage and revenue, would the company count the outcome as success? Would credible evidence of harm persuade it to restrict a profitable application?

The same test applies outside the chatbot. Would companies welcome religious advisers who insist that families should not shoulder the grid costs of serving data centers? Counsel about human dignity must reach corporate conduct as well as a model’s answers.

Religious participants must examine their own incentives, too. Access can be flattering. An adviser who fears losing influence may hesitate to criticize. A company could benefit from his reputation while declining his most consequential recommendations.

The available reporting does not establish a deliberate scheme to manipulate believers. It does justify asking whether religious involvement lends companies public credibility without giving advisers sufficient influence over decisions.

There is also a theological question. In July, I argued that AI requires Christians to revisit human identity and responsibility. We cannot enter corporate discussions as though those convictions were optional.

Genesis 1:27 grounds human dignity in being created in God’s image. Our worth does not depend on intelligence or productivity. Scripture’s account of human sin also gives us reason to scrutinize concentrated power, including our own exercise of influence. Religious consultation cannot substitute for accountability.

Teaching a system to produce language associated with compassion does not make it a pastor. Seeking its recommendation does not transfer our responsibility for a decision. Companies cannot acquire parental or spiritual authority merely by offering convenient services.

Cooperation across traditions remains worthwhile, but some disagreements about human nature cannot be reconciled. Companies should explain whose judgment prevails when advisers disagree. Christians should insist that their participation does not imply acceptance of the resulting decisions.

Companies should also report which consequential recommendations they adopt or reject. Participants should retain freedom to criticize and control how their names are used. Both should disclose relevant financial relationships, and companies should back claims of improved safety with evidence outsiders can examine.

I will stay engaged because AI offers real benefits and Christians owe a duty to the people it affects. Cootsona’s account suggests serious listening is possible. Companies should demonstrate what that listening changes, and religious advisers should remain willing to challenge decisions that violate their convictions.

“Faithful are the wounds of a friend” (Proverbs 27:6, ESV). Religious counsel earns its place by telling the truth, including when that truth requires a company to change course.