Tackling social media fraud with graph databases

The spread of misinformation presents a significant challenge in the civic space, and this last global “year of elections” has illustrated just how easily bad actors can slip through the cracks of social media platforms. Misleading claims fed to young TikTok users in the form of video montages proved to be endemic in the run-up to the UK election, with fictional, seemingly automated accounts behind them. Real or not, the content and subsequent comments may still succeed in convincing impressionable users with alternative facts.

In the U.S. 38% of adults used social media to seek information about the 2024 presidential election. With sweeping content moderation changes on Meta platforms imminent—leaving it up to users to comment on posts they deem to be inaccurate—manipulative information is increasingly likely to reach a broader segment of the population, and illegitimate social media campaigns will have the potential to mislead millions of people.

One way of avoiding further escalation means tackling the issue at the source. One way of doing this is through the use of graph database technology. This form of database structures and analyses information as entities and relationships and can help to untangle patterns and covert threads within them, ones which may seem legitimate but could be hiding scams or misinformation. By revealing these connections and delivering otherwise hidden insights, graphs can unleash the power of contextual data to fight all kinds of deception.

Here are three key qualities of graph technology that make it a powerful tool for investigating misinformation, fraud, and scams:

Recognizing relationships in data

Graph databases store data as a network of interconnected facts. This type of data model is useful for researchers to easily and quickly map and analyse complex connections. In the context of election misinformation, for example, it’s the relationships between social media ads, funders, and candidates that would hold insights.

By organizing the data as “nodes” and “relationships,” graph databases can enable researchers to surface hidden patterns and relationships between the ads and account credentials and then analyze those patterns and anomalies within the weakly connected components to discern malicious accounts.

Traverses relationships natively at scale and speed

Graph databases enable investigators to store detailed patterns of problematic actors. Then, they can query the data to uncover intricate connections between the suspicious actor and other entities.

By easily extending across data at scale and quickly identifying shared credentials between multiple accounts, analysts are able to spot areas for further investigation. Graphs easily encompass historical data, so users can uncover associations between different entities, like flagged and deleted social media accounts, for example, to build a more comprehensive analysis of how such networks can operate undetected on these social media platforms.

Uncovering massive financial fraud

It wasn’t just misinformation that was untangled by graph technology during this past crucial election year – it also uncovered previously hidden financial scams.

More than 3,000 entities that funded political advertisements on Facebook and Instagram, seemingly to influence voters, were identified by Syracuse University’s IDJC ElectionGraph Project. Most concerning was that the Institute found two entities connected to a complex network of bad actors who had disguised themselves as election campaigns to scam engaged voters out of money by promising merchandise like hats, flags, or coins in exchange for their credit card information.

Although these networks took steps to go undetected as they proliferated ads across Facebook, researchers used graph database technology to analyze the coordinated problematic content. They found some ads attempted to scam victims out of as much as $80 USD per month – and none of it directly supported any candidates’ campaign.

Keeping one step ahead with graph databases

Every year, organizations and consumers alike lose billions of dollars to online scams. Busting fraud and protecting users is all about finding and investigating the connections between various online entities, and modern technologies are helping to do so. Operating like a master detective, graph technology is capable of mapping patterns and relationships across huge amounts of data, enabling users to expose digitally savvy bad actors and helping them keep one step ahead of a complex challenge facing society.

Get yourself the best identity theft protection.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro