Adoption Patterns of Generative Artificial Intelligence in Healthcare Occupations: Cross-Sectional Study of User Interactions with Claude (Preprint)

Le résumé fourni par la source

BACKGROUND Generative artificial intelligence (GenAI) systems developed by organizations such as Anthropic (Claude), OpenAI (ChatGPT), Google (Gemini), Microsoft (Copilot), and Meta (Llama) are being rapidly integrated across various sectors, including healthcare. However, the extent to which GenAI is utilized for healthcare-related tasks in real-world settings, and its subsequent implications for patients and providers, remain largely unexplored. OBJECTIVE To quantify the frequency and scope of healthcare-related tasks performed using a state-of-the-art GenAI system (Anthropic's "Claude") and to evaluate the adoption patterns of this technology across different healthcare roles, considering usage by both healthcare professionals and the general public. METHODS This cross-sectional study analyzed anonymized data from over four million user interactions with Claude between December 2024 and January 2025. Interactions were pre-anonymized and classified by Anthropic's proprietary system into standardized occupational tasks. These tasks were subsequently mapped to specific healthcare activities using the U.S. Department of Labor's O*NET database. Main outcomes measured were: (1) the proportion of GenAI interactions by healthcare occupation, (2) the overall percentage of healthcare-related GenAI interactions compared to other fields, and (3) a "digital adoption rate," representing the proportion of tasks within healthcare occupations performed using GenAI. RESULTS Healthcare-related tasks accounted for 2.58% of total GenAI interactions. Among healthcare occupations analyzed, the highest interaction percentages were observed in Dietitians and Nutritionists (6.61%), Nurse Practitioners (5.63%), Music Therapists (4.54%), and Clinical Nurse Specialists (4.53%). Digital adoption rates across healthcare roles varied substantially, ranging from 13.33% to 65%, with an average adoption rate of 16.92%. CONCLUSIONS GenAI technology is experiencing selective yet growing integration into healthcare tasks, particularly in roles emphasizing patient interaction and education. This emerging pattern underscores critical considerations for the future impact of GenAI on clinical workflows, patient decision-making processes, and the reliability of healthcare information.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.