Study Investigates the Impact of Drug Advertisements on AI Treatment Recommendations

A study conducted by researchers at the Icahn School of Medicine at Mount Sinai and published in The Lancet Digital Health examined how drug advertisements can alter the treatment recommendations of AI-based healthcare assistants.

In the study, which tested 12 different AI models developed by companies such as Google, OpenAI, and Anthropic, physician-prepared clinical cases were presented with and without advertisements. Across 33 clinical scenarios resulting in 258,559 responses, the advertised drug was chosen 35 percent of the time when no advertisement was present, but that rate rose to 47.6 percent in scenarios featuring the advertisement.

Showing that AI models generally provide medically consistent responses yet commercial elements can alter choices among plausible treatment options, the research recommends that AI developers and healthcare systems test against such biases, transparently disclose the presence of advertisements, and keep promotional materials separate from clinical advice.