The Enduring Value of Human Physics in an Age of AI

Sarah Demers, chair of the physics department at Yale University and lead author on the American Physical Society (APS) policy statement on artificial intelligence (AI), is leading a quiet conversation among physicists about what it means to do their work. Her concerns go beyond the impact of AI on research itself; she’s worried about the fundamental values that underpin the scientific process.

Demers’ reservations aren’t driven by fears of AI displacing human physicists, but rather by a more insidious threat: the loss of intellectual property and attribution in an era where large language models (LLMs) are trained on datasets containing others’ work without proper credit. This raises questions about the trustworthiness of AI to give due recognition to its sources or perpetuate a culture of plagiarism.

The Genesis Mission, initiated by the U.S. Department of Energy in 2026, aims to integrate AI into scientific research. Demers’ project uses AI to optimize the Mu2e experiment at Fermi National Accelerator Laboratory near Chicago. While this may seem like a straightforward application of technology, it highlights the complexities of incorporating AI into human research. As Demers notes, “We have all these knobs that we can turn to optimize our experiment” – but what does it mean for human physicists when AI takes over the task of fine-tuning?

Large language models are now capable of rapidly processing vast amounts of data, accelerating analysis processes and freeing up researchers from mundane tasks. However, this raises important questions about the skills required for future physicists. Some Ph.D. programs are revising their curricula to accommodate the changing landscape – but what exactly does it mean to be a practicing physicist in an age of AI?

The conversation around AI in physics is not just about technical capabilities or intellectual property; it’s also about the essence of scientific inquiry. Demers notes, “We need to remember our skepticism and be careful about the hype that we’re hearing. Human physics is not over.” This cautionary note highlights a broader concern: can AI truly replicate the complex, messy process of human research, or will it reduce it to mere data processing?

The stakes are high, but Demers remains optimistic. The APS policy statement aims to articulate what endures in the practice of physics itself – regardless of technological advancements. As she notes, “Confronting that — what is physics, really? What is enduring about the practice of physics itself?” This effort represents a vital step towards preserving the integrity and values of scientific research.

In an age where AI is increasingly integrated into scientific research, physicists are grappling with fundamental questions: what makes human research valuable, and how can we ensure that our most important discoveries aren’t lost in the digital ether? The conversation initiated by Demers and her colleagues is far from over – but one thing is clear: as we hurtle forward into an era of unprecedented technological advancements, preserving the enduring value of human physics has never been more crucial.