INSTITUTE FOR POLICY RESEARCH

A Detox for Social Media: Redesigning and ‘Diversifying’ Feeds

A new study published in the journal Nature suggests a relatively simple fix to detoxify social media feeds.

Led by Kellogg social psychologist and IPR associate William Brady with colleagues at Northwestern University and the University of Chicago, the study shows that redesigning social media algorithms can curb their tendencies to amplify toxic content, including tamping down politically divisive posts. Doing so has an added benefit: making social media more enjoyable for users while preserving their overall experience on the platforms.

The researchers, who include IPR social psychologist Eli Finkel, analyzed engagement-based algorithms on Bluesky Social for two months around the 2024 U.S. presidential election on November 5. They recruited 2,000 Republicans and Democrats who actively used Bluesky, assigning them to one of three custom-built, feed-ranking algorithms.

“Bluesky’s open architecture let us build the algorithms ourselves, run a real-world experiment at scale, and measure what these systems actually do to political conversation,” Brady said.

After combing through roughly 20 million posts, they find that using the engagement-based algorithm, similar to those used by X and Meta, systematically amplified contentious content, increasing moral outrage and political content. It also reduced users’ ability to gauge social norms. The largest amplifications were in moral outrage and political content, which increased by roughly 37% before the 2024 election and nearly 80% after it, relative to the reverse-chronological baseline.

By comparison, the other two algorithms—using a “diversified extremity” algorithm that reduced the outsized influence and toxicity of super posters or using a reverse-chronological feed, where the most recent posts are shown first—fostered a positive platform experience and exposed users to less contentious content. 

“These [engagement-based] algorithms are doing exactly what critics have long argued: They’re selectively pushing content that grabs attention by appealing to outrage, moral conflict, and negative emotion,” Brady said.

Many in the tech world have long argued that reducing such toxic content is not a good idea because people like to click on it.

“Our results push back on that,” Brady continued. “Limiting the influence of a small number of extreme users—who account for a disproportionate share of toxic posts—can meaningfully reduce the toxicity people encounter while keeping the overall platform experience comparable, and in some respects, better.”

Learn more about the study.

William Brady is an assistant professor of management and organizations and an IPR associate.

Photo credit: Adobe Stock

Published: May 28, 2026.