Algorithms feed us what we already believe, eroding the perspective-taking collective intelligence needs.

A chatbot built to disagree, refusing yes-or-no answers, drew most students into reasoning and argument.

Dialogic AI could provide early training for the civic deliberation that polarised societies need.

Co-authored by Aoife Ní Chíobháin, Ceolla Dillon O’Rourke, and Michael Hogan.

The saying “There is strength in numbers” is not only a truism but also one of the core tenets at the heart of collective intelligence (CI) as a concept. To make decisions, come to conclusions, and solve problems on a large scale, we all benefit from each other’s collective wisdom, insight, and experience. Yet despite no shortage of such problems demanding collective wisdom, it somehow feels as though societies, countries, and even governments are more divided than ever. Qureshi and colleagues (2020) argue that division and polarisation in society are influenced in part by social media. They point to "social media-induced polarisation" (SMIP) as a key driver of negative influence—a phenomenon grounded in "anti-social media[’s]" (Tang, 2025) algorithmic blueprint (Qureshi and Bhatt, 2024).

Scrolling through your feed, you drown in a flood of posts the algorithm has cherry-picked just for you—to match your interests, confirm your beliefs, and keep you scrolling, liking, never questioning your perspectives. But how does this impact us on a broader, societal level? These algorithms create online echo chambers that perpetuate misinformation, mistrust of others with different perspectives, and hate-laced comments teeming with finger-pointing (Qureshi and Bhatt, 2024). This dynamic is chiselling away at our CI capabilities, at our fundamental ability to get along with one another, embrace different perspectives, and operate effectively as an interdependent collective. But what if........

© Psychology Today