The Landscape of Health Misinformation
A policy toolkit developed in partnership with the Rhode Island Public Health Association (RIPHA) as part of Brown's Watson School Master of Public Affairs program, with Sarah Carpenter, MD. The project applies a public-health prevention framework- tertiary, secondary, primary, and primordial- to the spread of health misinformation, tracing the problem from historical outbreaks like Ebola and Zika through the COVID-19 infodemic to the emerging risk of AI-generated health disinformation.
The work includes two Rhode Island-specific case studies (disinformation around the Reproductive Health Act and COVID-19 vaccine hesitancy), a review of current state legislation on AI use in healthcare, and concrete advocacy recommendations, including support for a Rhode Island media literacy curriculum mandate and legislative oversight of AI in insurance claims processing.
My contributions: researched Rhode Island's specific health misinformation landscape during the project's early scoping phase; authored the primordial prevention section, focusing on building institutional trust through policy opposing AI influence on health decision-making, and researched state legislation on AI use in healthcare across multiple states; co-authored the printed one-pager for the RIPHA symposium.
Prevention, layer by layer
The deepest layer, trust, is where the project argues AI policy belongs
Primordial
Build and maintain trust in health institutions
AI policy that protects patients, e.g. Rhode Island's AI Transparency Act (S2010/H7190)
Primary
Build health and digital literacy
Psychological inoculation · lateral reading · nudges
Secondary
Identify harmful narratives and intervene early
Pre-bunking · social listening (WHO's EARS) · filling information voids
Tertiary
Provide effective resources to combat false narratives
Debunking · community engagement to spread the truth
24.8%
of sampled COVID-19 tweets contained misinformation (Pasquetto, 2021)
102 in 65 min
vaccine and vaping disinformation blogs written by an LLM (Menz et al., 2024)