Most tools developed to counter the issues of misinformation and disinformation at scale replicate the very problem they aim to solve. By filtering, flagging, and moderating content, such tools replace human judgment with algorithmic judgment rather than empowering the people navigating the information ecosystem. This is what we refer to as the epistemic paternalism trap. We argue that AI tools designed to address misinformation should function as scaffolds for epistemic autonomy — explaining their reasoning and supporting individual judgment — rather than as opaque arbiters that decide on the user’s behalf. Three empirical and normative arguments converge on this conclusion. First, current algorithmic judgment is itself systematically biased and structurally less explainable than the human judgment it replaces. Second, correction-based ecosystemic interventions fail on two compounding grounds: they harm epistemic autonomy by substituting institutional verdicts for individual reasoning, and they fail on their own empirical terms, as their efficacy presupposes institutional trust that misinformation and disinformation have already eroded. Third, critical thinking education and transparency-oriented AI design better serve both epistemic autonomy and epistemic outcomes simultaneously — a convergence that holds even for those skeptical of autonomy’s intrinsic value. We argue that AI tools designed to address misinformation and disinformation should function as scaffolds for epistemic autonomy — explaining their reasoning and supporting individual judgment — rather than as opaque arbiters deciding what information people should or should not access.

Paper: https://doi.org/10.5281/zenodo.20321940

Automap: https://automap.borant.eu/share/h-V4VfSY5gTyxm6vzdKwt492Zf_aiTY3

Additional resources on the broader theme: https://www.giovannispitale.net/doing-research-with-ai/