Privacy, autonomy, and discrimination: perceived rights risks of AI in healthcare

Background

Healthcare AI is increasingly discussed as both a technical innovation and a governance challenge. Public concerns may influence trust, legitimacy, and acceptance of AI-enabled health systems. This study aimed to identify factors associated with perceived human rights risks of AI in healthcare among Internet users.

Methods

We analyzed data from a nationwide online survey of Internet users in Poland aged 18–65 years (N = 2,217). A quota-controlled sample was used to reflect the Internet-using population. Perceived AI-related human rights risk was assessed with five items covering key rights-related concerns about AI in healthcare. Items showed high internal consistency (Cronbach’s α = 0.94). Analyses were weighted to match the population structure. Multivariable linear regression was used to identify associated factors, including sociodemographics, e-health literacy (eHL), and general attitudes toward AI. The study was funded by the National Science Centre, Poland (Grant No UMO-2021/43/B/NZ7/02140).

Results

In the weighted regression model, the predictors explained 19.1% of the variance in perceived AI-related human-rights risk. Higher perceived risk was associated with greater fear of AI (B = 0.23; 95%CI 0.19 to 0.26), older age (B = 0.01; 95%CI 0.01 to 0.02), higher eHL (B = 0.13; 95%CI 0.02 to 0.23), and the highest Internet use (B = 0.29; 95%CI 0.10 to 0.48). Higher acceptance of AI was associated with lower perceived risk (B=-0.12; 95%CI -0.15 to -0.08). Selected education, income, and social media-use categories were also associated with perceived risk. Large language model use was not significant.

Conclusions

Public concerns about healthcare AI appear to reflect broader questions of trust, legitimacy, and perceived protection of individual rights, rather than exposure to AI tools alone. Responsible implementation should therefore combine technical safeguards with transparent governance that makes patient and citizen protection visible to the public.

Key messages

• Perceived human rights risks of healthcare AI were linked mainly to fear of AI and lower AI acceptance, rather than to social media or LLM use.

• Healthcare AI governance should make the protection of patients’ and citizens’ rights visible to strengthen public trust.