He doesn't believe in God. He just got a religious exemption from AI

David Durman did not have to believe in God to get out of using ChatGPT at work. He is a humanist, someone who builds a moral life without the supernatural. In September 2026, the University of Kentucky excused him from AI under the same law that gets a Seventh-day Adventist Saturdays off. Draft objectors of the 1960s laid the groundwork long before chatbots: they refused to fight without claiming any belief in God. Even earlier, in 1961, a Supreme Court footnote put “Secular Humanism” on the same list as Buddhism.

TechCrunch told the story on October 29, three days after the American Humanist Association announced the win. Another story shows what is at stake. In August 2025, Coinbase CEO Brian Armstrong described on a podcast how he gave his engineers a week to set up AI coding assistants. He called those who had not into a Saturday meeting, and some who had no good reason were fired. Armstrong himself called the approach heavy-handed. Durman’s case shows that workers have a legal way to say no.

A letter to the equal opportunity office

Durman is an administrator at the university. He helps researchers put together grant applications. Nobody has ordered him to use AI yet; colleagues and leadership have simply encouraged everyone to try the tools. He decided not to wait for the encouragement to turn into an order. Someone else’s example gave him the push: in the spring, software engineer Erin Maus had won the same kind of exemption.

In late August he wrote to the university’s Office of Equal Opportunity, asking for an accommodation based on his beliefs. He met with the office in mid-September, and on September 25 it approved the request, finding that it put no undue burden on his department. The timeline and the quotes from his letter come from a press release by the humanist association, which helped him file it.

His reasoning is philosophical. “The way we become moral actors for ourselves and for others is through the active application of our minds,” he told TechCrunch. Hand that thinking to a machine, he argues, and the capacity to act morally necessarily erodes. His letter lists other objections too: AI threatens people’s livelihoods, harms the environment and pushes out human creativity. It ends bluntly: “I cannot, in good conscience, partake in the use and proliferation of this technology.”

Religion without a god, legally speaking

In April 1964, Elliott Welsh filled out a conscientious objector form. He signed it, but only after striking out the words “my religious training and.” His views on war, he said, came from reading history and sociology. Welsh was convicted of refusing induction, but in 1970 the Supreme Court reversed the conviction. Deeply held moral convictions, the justices found, give the same right not to fight as faith does. Five years earlier, in Daniel Seeger’s case, the court had already extended that right to anyone whose belief holds a place in their life parallel to the one God holds for traditional believers.

The same logic runs through employment law. Title VII of the Civil Rights Act of 1964 bars employers from discriminating on the basis of religion and requires them to accommodate it. The Equal Employment Opportunity Commission (EEOC), the federal agency that enforces the law, defines religion very broadly. Belief in God is not required: moral beliefs about right and wrong count too, as long as a person holds them with the strength of religious faith.

The roots go back further still. In 1961 the Supreme Court heard the case of Roy Torcaso. Maryland had refused to commission him as a notary public because he would not declare a belief in God. The court sided with Torcaso. Footnote 11 of its opinion named religions that do not teach belief in God: Buddhism, Taoism, Ethical Culture and Secular Humanism. All of this is US law; other countries have their own rules.

Durman is not citing a fondness for old methods. He is citing a system of beliefs about where a person’s morality comes from, and that is what the law looks for.

Where belief ends and dislike begins

Telling your boss that ChatGPT annoys you will not get you an exemption. The EEOC says plainly that social, political or economic philosophies, along with mere personal preferences, are not religious beliefs. The law wants a whole worldview, not a product review.

An employer, however, has no business judging whether that worldview is true. Sincerity is generally presumed, and the commission notes that it does not matter whether someone holds a belief for the “proper” reasons. No church has to endorse it, either.

Erin Maus, the engineer who inspired Durman, shows how this works. In spring 2026 Maus, who works in North Carolina and is a Unitarian Universalist, asked to be excused from AI on ethical and environmental grounds. Her denomination has no official position on AI yet; one is still being drafted. In mid-May her company agreed. “I’m writing my code and reviewing my code by hand, which seems crazy to say,” she told Business Insider, as Futurism recounted. Without AI, she found, she finishes her tasks as fast as a colleague who uses it. AI, she said, does not really seem to change everything.

Why “no” is hard for an employer to say

Sincere belief alone does not guarantee an exemption: an employer can refuse if the accommodation would cost too much. In 2023, though, that escape hatch got much narrower.

Gerald Groff, a mail carrier, kept Sunday as a day of rest. Then the Postal Service signed a deal with Amazon and began delivering packages on Sundays. Groff quit in 2019, and his lawsuit eventually reached the Supreme Court. The ruling was unanimous, 9 to 0. Before that, many courts let employers refuse any accommodation that cost more than a trivial amount. Now the burden has to be substantial in the context of the business as a whole.

Applied to AI, that gets complicated. In June 2026, lawyers at the law firm FordHarrison laid out when a refusal might still hold up. The job may be impossible without AI, a client contract may require it, or it may be built into a core company system. General claims that AI is “important” or “efficient” may not be enough. Had one of the Coinbase engineers cited a sincere religious belief, the company would have had to show that excusing that engineer would seriously hurt the business. No appeals court has ruled on such a dispute yet; so far, every known case has been settled inside the company.

The AI button stays on the screen

Durman’s exemption has clear limits. According to FindLaw, a legal information site, he objected to tools such as Copilot in Microsoft Teams, ChatGPT and Claude, and he no longer has to use them himself. But he still attends meetings and submits materials through the same AI-equipped platforms as his colleagues. If a meeting runs on Teams, Copilot is still there; Durman just does not switch it on.

More requests like his are likely. Pope Leo XIV has called AI a threat to human dignity. Alexandria Huiskens, a staff attorney at the humanist association, says that concern runs through Catholic teaching, Unitarian Universalism and humanism alike. Her group wants nonbelievers to know that the law protects them too.

Durman says he is not recruiting anyone; he only hopes that people with similar doubts will feel free to act on them. The university said yes, with a condition. The accommodation works “while allowing the employee to continue performing the essential functions of their position,” university spokesperson Jay Blanton told TechCrunch. “As technology and job responsibilities evolve, accommodations may be reviewed to ensure they remain reasonable and workable.”