Information gerrymandering: Politicians forming divisive communities online

• Author:
• Author name

  Quantumrun Foresight
• October 24, 2022

Insight summary

With the increased use of social media, political views have become more divisive and combative. Many people seem to exist in partisan bubbles, physically and online. This one-sidedness is encouraged by political parties to keep their voters blind to the opposition’s views and policies.

Information gerrymandering context

Traditionally, gerrymandering manipulates electoral district boundaries to provide an unfair advantage to a constituency’s political party, group, or social class. This practice is often done through redistricting or re-allocating populations per district. In the US, segregated neighborhoods and racial voting patterns suggest that parties gain an advantage during redistricting by targeting communities of color. 

Gerrymandering is an old practice, but with computer algorithms and artificial intelligence (AI) advancements, map drawers can perform redistricting with even better precision, targeting well-defined voter demographics.

Alongside location gerrymandering, online exposure is also affecting voters’ subjectivity. Researchers call this information gerrymandering. In 2019, the Massachusetts Institute of Technology (MIT) conducted a study that placed participants in simulated gamified elections. The team found that communication networks (such as social media) can distort how others plan to vote and increase the chance of electoral deadlock or overall bias.

The researchers also created online bots, comprising about 20 percent of the total participants, to strongly support only one side, which the scholars called “zealots.” Over 2,500 volunteers participated in this study by playing the “voter game” under different conditions. After months of gameplay, the researchers discovered that election outcomes might be significantly influenced by how polling information was dispersed across networks and by the zealots’ activities.

Disruptive impact

Access to diverse information sources is crucial for informed decision-making in democratic processes. However, challenges arise when social networks restrict information flow or when information is distorted by biased individuals and automated bots. A study by researchers at the Massachusetts Institute of Technology (MIT) identified a phenomenon called information gerrymandering, where even without the presence of false information, the distribution of information can significantly skew group decisions. This phenomenon can create an electoral bias of up to 20 percent, leading to a situation where a group that should be evenly split 50-50 may end up divided 60-40 due to the unequal distribution of information.

The MIT researchers examined data on co-sponsored bills in the US Congress and European legislatures, as well as user networks on social media platforms. They found evidence of deliberate information manipulation to favor certain groups. This manipulation was evident in the analysis of co-sponsorship bills in the US from 1973 to 2007, where initially the Democratic Party exerted more influence. However, with the Republican Party's control of Congress in 1994, their influence balanced out with that of the Democrats. Similar patterns of polarization were observed in six out of eight European parliaments included in the study.

The findings of this research highlight the need to seek information from a variety of sources to form a well-rounded view, particularly in political decision-making. Companies, especially those involved in information technology and social media, may need to reassess their algorithms and policies to prevent information gerrymandering. Meanwhile, governments may need to develop regulations and guidelines to ensure a fair distribution of information, especially in politically sensitive environments. 

Implications of information gerrymandering

Wider implications of information gerrymandering may include: 

• Increased use of more subtle public surveillance technology to gather information about voters, such as facial scanning recognition and online activity.


• Community research groups collecting and providing unbiased information to their communities about candidates, policies, and more. 


• Increased use of zealot bots and troll farms to flood social media with often extremist ideals, which can result in real-world violence. 


• More computational propaganda campaigns from political parties to promote partisan ideals and spread false information against the opposition.


• AI increasingly identifing citizens likely to vote for a specific political party or support particular legislation.


• More vulnerable communities being targeted for voter manipulation or suppression.

Questions to consider

• What are some examples of information gerrymandering you have encountered?


• How else can information gerrymandering affect local communities?