Un employé du Kentucky obtient une exemption religieuse de l'utilisation de l'IA au travail

David Durman, a research administrator at the University of Kentucky, has secured a legal religious exemption that allows him to refrain from using artificial intelligence in his professional duties. The university approved the accommodation on September 25, marking the first known instance of a workplace AI accommodation granted to a humanist. Durman, who identifies as a nontheistic Humanist, argued that employing generative AI conflicts with his core beliefs regarding human dignity and intellectual autonomy. While he is not required to personally utilize the technology for tasks such as editing grant applications or handling compliance matters, the ruling permits him to attend meetings involving AI and collaborate with colleagues who use the tools. The path to this exemption began in late August when Durman contacted the American Humanist Association, a nonprofit organization dedicated to advancing secular humanism. He sought guidance on securing legal protections against mandatory AI usage, a concern that grew as the university began encouraging staff to integrate the technology into their workflows. Although Durman noted that he had not yet been forced to use AI, he viewed the request as a preemptive measure. He feared that recommendations to adopt these tools could evolve into mandates in the future. After meeting with the university’s Office of Equal Opportunity in mid-September, the administration determined that granting the accommodation would not impose an undue burden on his department. Durman, 45, explained that his humanist beliefs center on human well-being and flourishing, both individually and collectively. He views the active application of the mind as essential to becoming a moral actor. Consequently, he argues that relying on AI diminishes critical thinking skills and negatively impacts an individual’s ability to reason independently. In a statement released by the AHA, Durman highlighted broader societal concerns, stating that AI interferes with livelihoods, poses environmental risks, and threatens human creativity in arts and literature. He emphasized that protecting the capacity to think and reason for oneself is paramount to his lifestyle. This case follows a similar precedent set by Erin Maus, a Unitarian Universalist software engineer based in North Carolina. Maus secured comparable protections in May after raising religious, environmental, and ethical objections to the technology. Her case is considered particularly unusual given the rarity of software engineers writing code without AI assistance. Durman cited Maus’s success as inspiration for pursuing his own accommodation. Both cases illustrate how existing workplace religious protections are being tested as employers increasingly embrace generative AI and some employees raise moral or philosophical objections to its use. Alexandria Huiskens, a staff attorney at the AHA, noted that concerns about AI and human dignity span multiple faiths and traditions, including Catholicism, Universalism, and Humanism. She stated that the organization wants nontheists to understand that they are protected under current laws regarding workplace accommodations. Durman hopes his public discussion of the process will encourage others with similar convictions to seek resolutions appropriate to their situations. He clarified that he is not attempting to dictate how others should think or feel, but rather advocating for the right to pursue personal concerns within the workplace. The University of Kentucky spokesperson, Jay Blanton, confirmed that the institution reviews such requests on an individual basis, taking into account specific job responsibilities and circumstances. Blanton stated that the university determined an accommodation could be provided while ensuring Durman continued to perform the essential functions of his position. He added that as technology and job roles evolve, such accommodations may be reviewed to ensure they remain reasonable and workable. The approval coincides with a rapid expansion of AI integration across American workplaces, including within the University of Kentucky itself. The university has implemented the Commonwealth AI Transdisciplinary Strategy, or CATS AI, a framework designed to enhance AI literacy and tool accessibility for students, faculty, and staff. By June, the university’s AI training hub had connected more than 2,000 employees across 17 colleges, centers, and 50 departments with learning opportunities. Despite this institutional push toward adoption, Durman’s exemption stands as a distinct carve-out for individual belief systems. Durman distinguished his objection specifically from general technological use. He drew a line between passive tools like autocorrect and generative AI or large language models, citing the latter’s potential to facilitate the offloading of cognitive abilities and decision-making processes. His approach to the accommodation process was described as straightforward, taking approximately one month from initial contact with the AHA to final approval. As companies navigate the ethical and practical implications of integrating AI into daily operations, cases like Durman’s offer early insights into how labor laws intersect with emerging technologies and diverse philosophical frameworks.