Fake news remains a persistent challenge, particularly during election seasons when conspiracy theories and misinformation from malicious sources seek to sway voters. As the US election heats up in one of the tightest races yet, researchers at Ben-Gurion University of the Negev have created a method to assist fact-checkers in managing the surge of misinformation on social media. Led by Dr. Nir Grinberg and Prof. Rami Puzis, the team discovered that by tracking the sources of fake news rather than focusing on individual articles or posts, they can significantly reduce the workload for fact-checkers while producing reliable results over time.
Dr. Grinberg explains, “The issue today is that fact-checkers are inundated with content and struggle to evaluate everything. Their coverage amidst the vast array of social media activity is often unclear, and we lack insights into how effective they are at prioritizing the most critical content. This inspired us to develop a machine learning method to help fact-checkers optimize their focus and enhance their productivity.”
Their findings were published in the Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining.
Given the rapid rise and fall of fake news sources, keeping updated lists can be costly and labor-intensive. The researchers’ system takes into account the flow of information on social media and the audience’s tendency to engage with falsehoods, making it more robust and effective over time. Their audience-based models outperformed the traditional method of tracking who shares misinformation, achieving 33% better accuracy with historical data and 69% better when identifying emerging sources. Notably, their approach maintains high accuracy in identifying fake news sources while reducing fact-checking costs by over 75%.
While the system requires further training in real-world scenarios and should not replace human fact-checkers, Dr. Grinberg emphasizes that it can significantly enhance the coverage of current fact-checking efforts. Both he and Prof. Puzis are part of the Department of Software and Information Systems Engineering.
As Grinberg and his team work to improve the integrity of elections, a crucial question remains: will social media platforms take the necessary steps to provide data and access to combat misinformation effectively?
The research team also included Maor Reuben from the Department of Software and Information Systems Engineering at BGU and independent researcher Lisa Friedland.
Author: shadjavist
Humble gentleman who never believes in impossibility
