Research Summary: Breaking News Out of the Filter Bubble: Generative AI Search Diversifies Collective Attention and Raises Shared Information Consumption

A large-scale randomized field experiment with The Washington Post readers indicates that generative AI search, including AI overviews with article citations, significantly diversifies information consumption. Contrary to concerns about filter bubbles, this technology expands the reach of popular topics and increases overlap in reader attention, while simultaneously reducing concentration on highly popular topics and shifting consumption towards less-popular content, both for individual readers and across the broader audience.

Why it matters

This research provides crucial insights into the evolving landscape of information dissemination and consumption, particularly with the rise of generative AI in search. It challenges prevailing assumptions about AI's potential to narrow information exposure, suggesting instead that it may foster a more diverse and shared understanding of information across populations. Understanding these dynamics is critical for shaping future digital strategies and platform design.

Key insights

• Generative AI search expands the reach of widely read topics.
• AI search increases the overlap in topics consumed by readers.
• Consumption of information becomes less concentrated on popular topics when using AI search.
• AI search shifts consumption towards less-popular topics.
• These effects are observed both within individual readers' consumption patterns and across the entire audience.
• The study was a randomized field experiment involving 37,561 readers at The Washington Post.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2609.38946