Uncovering misinformation Knowledge graphs help in the 2024 US election campaign

Related Vendors

neo4j

FAULHABER Drive Systems

To uncover misinformation in the U.S. election campaign, a project examines political advertisements and posts on social media. The data is modeled and analyzed in a knowledge graph.

False and inaccurate information can have devastating effects on the outcome of an election. False information was circulated primarily in closed groups and communities on social media.

The Institute for Democracy, Journalism and Citizenship (IDJC) at Syracuse University in the US has published initial results from a one-year research project on the 2024 US Presidential election with the IDJC ElectionGraph. The report takes a look at political campaigns and advertising on social media platforms.

Knowledge graph to detect misinformation.

The goal is to identify actors who spread false information (fake news) and try to influence US voters using a knowledge graph. The research project is, among other things, supported by Neo4j. The provider of graph databases and graph analytics supports the research project with 250,000 US dollars and free licenses of its graph database.

For the report, the research team led by Professors Jennifer Stromer-Galley, Margaret Talev and Johanna Dunaway examined mentions of presidential candidates Joe Biden and Donald Trump from September 2023 to February 2024. The team focused on the Facebook and Instagram platforms, which together reach more than half of the US electorate.

Artificial Intelligence in the Election Campaign

The key questions include: Who are the most influential players? What topics are circulating on which platforms? What misinformation is being spread and where (for example, swing states)? What effect do they have? And what role does AI play?

"The results of the IDJC ElectionGraph give us an insight into the flood of information and misinformation from groups with different motives, connections, and credibility that are raining down on voters ahead of the 2024 elections," says Margaret Talev, Kramer Director at the Institute for Democracy, Journalism and Citizenship (IDJC) at Syracuse University.

"The graph algorithms of Neo4j help to uncover connections and interactions in a complicated network and to identify the sources of these messages."

Graph algorithms sift through information

Neo4j has not helped journalists to make political connections and financial transactions transparent for the first time. In 2015, the International Consortium of Journalists (ICIJ) used the graph database for the analysis of the Panama Papers, thereby uncovering one of the biggest global corruption scandals of all time. Neo4j was also used as an analysis tool for the Paradise Papers in 2016, the FinCEN Files in 2020, and the Pandora Papers in 2021.

AI-generated information is increasing

"For data scientists and data journalists, it's anything but easy to keep up with the mass of AI-generated and disseminated fake news," says Jim Webber, Chief Scientist at Neo4j. "Graph technology can be a powerful tool for uncovering hidden patterns and networks in large amounts of data and for exposing attempts at manipulating democratic elections, for example."

"At Neo4j, we are proud to support Syracuse University's ElectionGraph and thereby contribute a little to ensuring that voters can make their decision based on objective facts and information."

"The report of Syracuse University's "IDJC ElectionGraph" is available for free download."