When discussing any topic online, there often seems to be two distinct groups with conflicting opinions. This division is mainly influenced by social networking platforms’ algorithms, which connect users with like-minded individuals. These platforms can create echo chambers that worsen polarization.

The susceptibility of these platforms to external manipulation makes them attractive to malicious actors aiming to disrupt societies. A recent study conducted by Concordia researchers and published in the journal IEEE Xplore introduces a novel method of maximizing polarization through social media accounts using artificial intelligence.

The lead author of the study, Mohamed Zareer, emphasizes the importance of enhancing detection mechanisms and identifying vulnerabilities in social network platforms to combat malicious manipulation.

A little data can do a lot of harm

The researchers utilized data from millions of Twitter accounts related to vaccine opinions and employed Double Deep Q-Learning to create adversarial agents. These agents, equipped with limited information, exhibited significant potential in intensifying polarization across social networks.

The ultimate goal of this research is to prompt policymakers and platform owners to implement new security measures against malicious agents and ensure responsible AI usage.