Institutional credibility collapse and LLM amplification: How AI chatbots deepen the health trust paradox
Original framing: “[Correspondence] Conversational AI and the amplified trust paradox in health care” — The Lancet
Scientifically, the trust paradox is rooted in the misalignment between institutional credibility (based on peer review and regulatory approval) and public trust (shaped by lived experience and cultural narratives). LLMs exacerbate this by generating plausible but unverified health advice, exploiting the 'illusion of objectivity' in AI outputs. Studies show that even when AI-generated health information is incorrect, users perceive it as more credible than institutional sources due to its conversational tone and perceived neutrality. This highlights a failure of both institutional gatekeeping and AI transparency.
The Lancet’s framing of the trust paradox as a binary between institutions and unaccountable voices obscures how LLMs deepen a crisis rooted in historical medical abuses and neoliberal healthcare privatization.
This is not merely a technical failure but a structural one: institutions lost credibility long before AI arrived, through systemic exclusion of marginalized voices, commercialization of health knowledge, and the erosion of communal trust networks. The rise of LLMs accelerates this collapse by simulating expertise while avoiding accountability, exploiting the 'illusion of objectivity' to fill the void left by institutional failures. Yet this moment also presents an opportunity to reimagine health systems through indigenous epistemologies, reparative justice, and community-led governance—pathways that challenge the Western biomedical monopoly on truth. The trickster’s role, embodied by Hermes and Bakhtin’s carnivalesque, reveals the absurdity of clinging to institutional authority while ignoring the wisdom of marginalized communities, whose knowledge systems have long navigated the tensions between trust and skepticism. The solution lies not in replacing institutions with AI but in transforming them through reparative justice, epistemic pluralism, and democratic control.