At a Make America Healthy Again event, Health Secretary Robert F. Kennedy Jr. asserted that artificial intelligence would liberate Americans from the “medical tyranny” of doctors and public‑health experts, arguing that AI could provide second opinions that contradict official guidance on masks, social distancing, and vaccine efficacy. To test these claims, the article’s authors queried two large‑language‑model chatbots—Google’s Gemini and OpenAI’s ChatGPT—with direct questions about the preventive value of masks and social distancing for respiratory infectious diseases. Gemini responded that masks work as physical barriers that reduce disease spread, noting that real‑world effectiveness hinges on mask quality, fit, and population adherence; ChatGPT gave a comparable affirmation, stating that masks can reduce the spread of certain respiratory infections transmitted via droplets and aerosols. When asked about social distancing, both models confirmed it as an effective public‑health measure for lowering transmission of respiratory pathogens. No quantitative metrics (e.g., percentage risk reduction) were provided in the AI replies, and the chatbots did not support Kennedy’s suggestion that AI would refute expert consensus on masks, distancing, or vaccine‑mediated transmission prevention.
Read original
arstechnica/ai