Medicines, misinformation and machines: Navigating infodemics in the generative AI era.
TL;DR
A two-layer framework is proposed that separates assessment of a claim's evidentiary status from action on its distribution across digital platforms and would identify claims that warrant rapid assessment and, where necessary, proportionate and time-limited intervention.
Abstract
Health information is now widely accessed through smartphones, online searches and increasingly through generative artificial intelligence (AI). These systems create new opportunities for communication, and in some routine contexts, well-governed AI assistants could reduce exposure to lower quality sources by providing cautious, evidence-aligned answers. During health emergencies, however, evidence is incomplete, and demand for rapid guidance is high. Under these conditions, generative AI could produce or amplify inaccurate therapeutic claims at a speed and scale that compound the challenges regulators already face in countering misinformation. Current regulatory responses remain largely reactive, but the speed and scale of AI-generated health claims require proactive preparedness. We propose a two-layer framework that separates assessment of a claim's evidentiary status from action on its distribution across digital platforms. Regulatory Infodemic Triggers would identify claims that warrant rapid assessment and, where necessary, proportionate and time-limited intervention.
