The recent US elections have brought forward allegations of fraud with some stakeholders, both within and outside the US, using Benford's Law to support their allegations. We are pleased to invite you to a webinar hosted by the School for Data Science and Computational Thinking. Prof Golbeck will talk on Benford's Law and its applicability to fraud. Recently, she was featured on Episode 4 of "Connected" on Netflix talking about Benford's Law. This talk, as with all our webinars, is open to all. Please feel free to pass this information about the webinar to others who may be interested, including students and interested parties outside the university.
Title: Using Benford’s Law to detect bots, social media fraud, but maybe not election fraud
Speakers: Jen Golbeck (University of Maryland)
When: Thursday 12 November, 14h00 - 15h00 (SAST)
Where: Please email firstname.lastname@example.org for a Zoom link.
Benford’s Law is a statistical rule that says, in some systems, the distribution of first significant digits (the first digit in a number) follows a known pattern. Numbers starting with 1 occur around 30% of the time, decreasing to numbers beginning with 9 only occurring around 5% of the time. This rule is so reliable that it can be used in court as evidence of financial fraud. Our work has shown it can also help detect bots and like fraud on social media. More recently, questions have been asked by some Trump supporters in the US about using Benford’s Law to detect election fraud, but the evidence for that is far more mixed.
Jennifer Golbeck is a Professor in the College of Information Studies at the University of Maryland. Her research focuses on artificial intelligence and social media, privacy, malicious social media behaviour, and trust on the web. She has a PhD in computer science from the University of Maryland and degrees in computer science and economics from the University of Chicago.