The next speaker in this session at the 2026 Association of Internet Researchers conference in Mexico City is Michelle Kulig, whose focus is on perceptions of algorithmic recommender systems amongst news avoiders in the United States. The US political environment is characterised by terminal polarisation, with competitive news markets that incentivise outrage while being subjected to political pressures, and an absence of strong and trusted public service media that might function as an epistemic anchor.
News consumption here is now primarily via social media environments, which are to some extent algorithmically curated and therefore produce further epistemic disruption. There is low trust in the reliability and accuracy of news reporting, and a substantial amount of intentional news avoidance, for a variety of reasons (news overload, distrust, negative impacts on wellbeing). This is a particular feature of US news contexts, even in comparison to other western democracies.
How might we tackle such epistemic disruption and improve epistemic welfare, then? What role can knowledge-bearing institutions and practices, algorithmic recommender systems, and other components of the news environment play here? Much of the focus to date has been on the role of journalism here, and less so on algorithmic information infrastructures and user understandings of such systems. Can changes to user evaluations of algorithmic recommender systems help here?
This study explored this through a representative survey of some 1,400 US adults, with a particular focus on intentional news avoiders. It explored their motives for such avoidance, asked them to evaluate recommender systems, and discuss their design preferences for such systems.
60% of respondents identified as intentional news avoiders, due to negative mood impacts (60%), news distrust (47%), news overload (42%), and feelings of powerlessness (42%). Their evaluations were slightly less favourable than others’, and their avoidance motives shaped such perspectives – this was true especially for distrust-driven avoiders. News avoiders also expressed a strong demand for transparency and control features in such systems: distrusting users wanted to know and alter data use, and alter the algorithm; others wanted more transparency, alter data use, and opt out altogether.
US news avoidance goes beyond content dissatisfaction, then: it is a structurally embedded response to epistemic disruption, and increasingly also targets algorithmic information infrastructure. To mitigate this, affective harm and distrust must be addressed.
