As the University of Kentucky began encouraging employees to use AI—from Copilot for meeting notes to tools like Claude and ChatGPT for everyday tasks—David Durman, a research administrator, worried about what he would lose by outsourcing his own thinking to machines.
So in August, Durman, who identifies as a humanist, took an unusual step. He sought a religious exemption from using artificial intelligence at work. With help from the American Humanist Association, a nonprofit that advocates for the rights of humanists and atheists, he submitted a letter to the university arguing that relying on AI to perform tasks that require human judgment, listening, and personal responsibility violated his deeply held beliefs. Doing so, he wrote, would diminish human dignity and amount to an abdication of “human moral agency.”
On September 25, the university granted his request, determining that the accommodation would not impose an undue burden on his department.
The decision offers a potentially powerful new avenue for workers who object to AI, allowing them to invoke federal religious-accommodation protections to opt out of a technology that employers are increasingly embracing. And as Durman’s case illustrates, workers don’t necessarily have to believe in God to qualify.
Why humanists are pushing back on AI
Critical thinking is a core tenet of humanism, a nontheistic belief system that emphasizes human dignity, reason, and our responsibility to care for one another. Last year, an international coalition of humanist groups adopted a declaration outlining principles for AI, including that the technology should never replace human judgment and should benefit society broadly rather than enrich a privileged few.
There’s no single humanist viewpoint about exactly how to use AI, or whether it should be used at all. “You can be a humanist and believe different things about AI,” says Fish Stark, executive director of the American Humanist Association, who left the tech industry himself because of his own concerns about the risks of AI. What unites humanists, he says, is a belief that the technology should advance human flourishing, not undermine it. “We think the thing that makes people special isn’t an immortal soul; it’s conscience and the ability to think for ourselves,” Stark says.
In a press release, Durman also cited his concern about the broader impacts of AI, from the potential for job loss to environmental damage—also things that obviously conflict with his values as a humanist.
Under Title VII of the Civil Rights Act, employers generally must accommodate employees’ sincerely held religious beliefs, including certain nontheistic beliefs, unless doing so would impose an undue hardship on the business. While accommodations for religious observances, such as not working on the Sabbath, are well established, requests to avoid using AI are of course a much newer development.
Durman isn’t the only employee who has successfully taken this approach. Earlier this year, a software engineer who is a Unitarian Universalist also won an exemption from using AI at work, citing her concerns with the environmental impact of AI use. After Pope Leo XIV wrote about the potential risks of AI—including the potential for loss of human judgment and independent thinking—some Catholics may also begin seeking exemptions.
A new way to opt out
The American Humanist Association is now encouraging others to follow Durman’s example. On its website, the organization has posted an open invitation offering legal assistance to humanists interested in seeking workplace AI exemptions. “In a world where people need and don’t have more control over the development of AI, one of the few tools that we have is the ability to withdraw our consent and our participation,” Stark says.
You don’t need to have officially called yourself a humanist in the past to be able to potentially take this route. “Humanism isn’t necessarily about showing up to a church and practicing; instead it’s about what you believe and what you do in your life,” Stark says. “So we think probably about 10%-15% of American adults are humanists and many of them don’t use the term. To attain a religious accommodation, the standard is that you have to show that you have a sincere and deeply held belief in that religious tradition. Whether you’ve called yourself a humanist for a long time [or not], if you’ve believed what humanists believe and you believe it deeply, then that meets the standard.”
In religious accommodation cases, employees need to prove that their objection is religious and not political or coming from some other motivation, says Dallan Flake, a professor at Gonzaga University who studies employment law. If the objection is religious, employers have to offer the accommodation as long as it doesn’t cause the company or organization undue hardship. In the past, it was fairly easy for employers to claim that any hardship was too much, but that changed with a 2023 Supreme Court case. “It’s a much higher burden,” says Flake. Now, if an employer wanted to prove that an employee avoiding AI would cause them hardship, they’d have to show that the change caused substantial increased cost or loss of productivity.
For most jobs, that’s probably difficult to prove—especially since it hasn’t been long that AI was in widespread use, and people have been able to do their work without it. “I do think in the future, as we become more dependent on AI and it becomes more integrated into our job duties and as employers expect more fluency in using AI, at that point they might have a stronger claim that, hey, if we exempt this employee from using AI, that’s really going to harm us,” Flake says.
From a humanist perspective, beyond the choices that people make about their own work, Stark argues that those who are concerned about AI should be pushing for collective action. “I think that people should be coming together to demand better AI governance from our government,” he says. “Ultimately AI has the potential to do tremendous good or tremendous harm to humanity.”
Its development, he argues, shouldn’t be controlled by “a small cabal of billionaires.” Instead, it should be subject to democratic oversight.