UTSA cyber researcher building tools to weed out dishonesty online
Raymond Choo, a former police officer in Singapore turned cybersecurity research professor at the University of Texas at San Antonio, is developing software to detect dishonesty online by running software that scans the internet for similar writing patterns.
The pilot software hones in on word choices, punctuation and context and is able to discern whether multiple comments came from the same source. Sometimes marketers use fake social media accounts to bolster the ranking of businesses on social media and websites that have public reviews of products and services.
As more businesses collect data about how customers interact with companies online, and often dedicate resources to respond to them on social media networks, this type of tool could be used to parse out trolls or artificial intelligence powered posts.
The initial process of this recent study uses n-gram, or a sequence of letters, that can be parsed through for analysis.
“In the initial research we tried to search using psychometric profiles or methods that could be used in intrusion detection and through the course of that research we discovered that — even though sometimes we try to disguise ourselves — our writing gives us away,” said Choo, the new associate professor of Information Systems and Cyber Security at UTSA
The goal is to eventually commercialize the software, but the race against threat actors, or individuals with malicious intent, to develop something that can outlast their techniques remains challenging.
“In the longer term, we are not going to restrict ourselves to n-gram,” said Choo. “We are in the process of fine tuning a few statistical algorithms into something that we can use.”