Large networks of rational agents form persistent echo chambers

Abstract

Echo chambers (ECs) are enclosed epistemic circles where like-minded people communicate and reinforce pre-existing beliefs. It remains unclear if cognitive errors are necessarily required for ECs to emerge, and then how ECs are able to persist in networks with available contrary information. We show that ECs can theoretically emerge amongst error-free Bayesian agents, and that larger networks encourage rather than ameliorate EC growth. This suggests that the network structure itself contributes to echo chamber formation. While cognitive and social biases might exacerbate EC emergence, they are not necessary conditions. In line with this, we test stylized interventions to reduce EC formation, finding that system-wide truthful ‘educational’ broadcasts ameliorate the effect, but do not remove it entirely. Such interventions are shown to be more effective on agents newer to the network. Critically, this work serves as a formal argument for the responsibility of system architects in mitigating EC formation and retention.

Influence and seepage: An evidence-resistant minority can affect public opinion and scientific belief formation

“…We design our agents to be Bayesian not only because people's decisions can conform to Bayesian norms of rationality (e.g., Griffiths, Kemp, & Tenenbaum, 2008;Lewandowsky, Griffiths, & Kalish, 2009), but in particular because even seemingly "irrational" behaviors can emerge from Bayesian principles. For example, belief polarization Jern, Chang, & Kemp, 2009) can be accommodated within a rational Bayesian framework, and it has been shown that Bayesian agents can form persistent "echo chambers," enclosed epistemic bubbles in which agents share most beliefs (Madsen, Bailey, & Pilditch, 2018). The use of rational agents also seemed advisable in light of several suggestions that climate denial can be considered a rational enterprise Lewandowsky, Cook, & Lloyd, 2016), notwithstanding its wholesale dismissal of scientific evidence.…”

Analytic Versus Computational Cognitive Models: Agent-Based Modeling as a Tool in Cognitive Sciences

“…Recently, a number of such simulations and computational models have been used to shed light on important issues in modern societies. For example, using computational cognitive models of belief revision in social networks, we may explore belief diffusion (Duggins, 2017), emergence of echo chambers (Madsen, Bailey, & Pilditch, 2018), belief cascading (Pilditch, 2017), and network pruning (Ngampruetikorn & Stephens, 2016).…”

“…Whereas we may develop models to deal with isolated cognitive functions (e.g., memory retrieval, belief revision), the act of implementing models in relatively large social systems is useful in explicitly accounting for all aspects of the social system. For example, belief-revision models might require information-search strategies when dealing with social networks (see Madsen et al, 2018), and the nature of these strategies is made explicit in attempts to model them.…”

Complex sociotechnical systems as a subject of constructivist epistemology research: a descriptor of subjectivity

“…Wagner, Paul M., and Tuomas Ylä-Anttila are among those researchers who believe that people and organizations tend to get information from those whose beliefs are similar to their own, forming "echo chambers" with their network connections. The effect of echo chambers has also been compared to epistemic circles and networks (Madsen, 2018), where like-minded people actively communicate and become increasingly entrenched in their beliefs.…”