Algorithms and the Value of Content

When we engage with social media, we expect to see content that reflects our values and interests. However, new research has revealed that social media algorithms often prioritize content that clashes with our core beliefs.

Researchers from the Massachusetts Institute of Technology (MIT), the University of Michigan, and Stanford University conducted a study on the X social media platform to understand how algorithms select content for users’ feeds. They analyzed the feeds of 715 U.S.-based users and found that the algorithm favors content that contradicts users’ values.

The study used a measurement tool to classify the values expressed in the posts users saw in their feeds. The researchers found that the X feed algorithm is more likely to amplify posts about upholding tradition, following rules, or keeping society safe, while demoting posts about looking after people, concern for people far away, being dependable, or protecting nature.

When compared to the values users expressed in their own posts, the algorithm was more likely to promote content that conflicted with users’ values than content that aligned with them.

So, why does this happen? The researchers found that users primarily respond to content by liking it, but the algorithm treats replies as a stronger signal. As a result, the algorithm tends to send users posts that reflect the values of posts they’ve commented on, which often clash with their own values.

Notably, the study found that this algorithmic tendency is stronger for users who reported being Democrats than for those who reported being Republicans. The researchers found that Democrats are more likely to object to content they reply to, creating a stronger feedback loop that amplifies clashing posts.

The study suggests that social media platforms can better align feeds with users’ values by asking users what they value and sorting their feeds accordingly. This approach may help reduce polarization and foster constructive conversation.

Ultimately, the researchers argue that users should have a greater say in the kinds of information shown to them. If social media platforms can create new tools that facilitate this goal, they may be able to better support the values and actions people care about.

Why Do Social Media Algorithms Clash with Our Values?

According to the researchers, the primary reason social media algorithms clash with our values is the way users interact with content. While users tend to like posts that reflect their values, they rarely write replies to those posts. However, when users do reply, they often respond to content that clashes with their values.

The algorithm treats these replies as a stronger signal than likes, which means it learns most strongly from them. As a result, the algorithm tends to send users posts that reflect the values of posts they’ve commented on, which often clash with their own values.

This phenomenon is particularly pronounced for users who reported being Democrats. The researchers found that Democrats are more likely to object to content they reply to, creating a stronger feedback loop that amplifies clashing posts.

The study suggests that social media platforms can mitigate this issue by asking users what they value and sorting their feeds accordingly. This approach may help reduce polarization and foster constructive conversation.