AI Tool Reduces Political Polarization on Social Media

AI Tool Helps Users Avoid Political Polarization on Social Media

Researchers have found a way to lessen political polarization caused by social media algorithms. They created a browser extension that changes how users see posts on X, the platform formerly known as Twitter. This tool helps show that algorithms can lead to extreme views, but it also offers a way to reduce this effect without needing permission from the platform owners.

Understanding Political Polarization and Algorithms

Many people think social media algorithms make political views more extreme. However, proving this has been tough since only the platform owners can see how their algorithms work. Researchers often struggle to study these algorithms without help from the companies behind them.

New Study Shows Clear Evidence

A recent study published in Science provides strong proof that algorithms cause polarization. The researchers created a tool that can reorder posts in a user’s feed. This helps to lower the visibility of polarizing content, such as posts that promote hostility or support for undemocratic actions. The tool uses a large language model to analyze and adjust the posts in real-time.

“Only the platforms have had the power to shape and understand these algorithms,” says Martin Saveski, a co-author of the study. “This tool gives that power to independent researchers.”

Impact of the Browser Extension on Users

The researchers ran an experiment with over 1,200 volunteers leading up to the 2024 U.S. election. Participants were shown feeds where polarizing content was either pushed down or slightly boosted. Those who saw less polarizing content reported feeling more positive towards people with opposing views.

Measuring Changes in Feelings

The results showed a change in feelings by two to three degrees on a 100-degree scale. This might seem small, but it’s similar to three years of change in average feelings in the U.S. Those exposed to less polarizing content reported feeling less sadness and anger while scrolling through their feeds.

Future Research and Tool Development

Experts believe the findings from this study are significant. Victoria Oldemburgo de Mello, a psychologist, praised the study for its cleverness and real-world application. The long-term effects of using this tool are still unclear, but the researchers plan to keep studying it. They have also made their code available for other scientists to use, which could help in further research.

Challenges and Future Directions

The current version of the tool works only on web browsers. Researchers are looking into making a version for apps, although that would be more complicated. They also hope to explore other ways to adjust social media feeds, focusing on well-being and mental health.

“Our framework is very general, and one can think about well-being, mental health, and so on,” says Saveski.

Conclusion: A Step Towards Better Social Media

This research shows that it is possible to reduce political polarization on social media through simple adjustments. The tool created by the researchers opens up new possibilities for understanding how content affects users. In the future, we might see more tools like this that help make social media a better space for everyone.