Healthcare has a non-clinical problem shaping what reaches the bedside. A 16.4-point rise in awareness of generative AI in medicine between May 2023 and November 2025 bought essentially no movement in comfort with AI-led primary care or AI-led therapy, and the only demographic axis that shifted in that window was the one clinical evidence cannot address: party.
Rader, Gertz, Cantor, McBain, and Brownstein's two-wave survey on the Outbreaks Near Me platform (N = 3,130 in 2023, N = 3,088 in 2025) put a number on the move: across four comfort and perception questions, interaction odds ratios between political identity and the trend ran 1.69 to 2.11, with Republicans trending more favorable and Democrats declining. The mechanism is older than the technology. New medical tools have historically been accepted when they bend to existing clinical relationships; the gap between the two parties on whether genAI belongs in those relationships is not closing with exposure.
That changes the equity math. If comfort tracks political identity, deployment plans cannot assume acceptance follows availability, and the populations least likely to adopt will be the ones a rollout reaches first. Equitable deployment now has a non-clinical variable to plan around, and clinical evidence alone will not move it.