Until now, most disinformation research, datasets, and tools to protect users like bot detectors, have only included English-language social media. This lack of linguistic diversity leaves Francophone internet users in Quebec and beyond more vulnerable to disinformation as the phenomenon becomes increasingly pervasive in online spaces.
A new research project funded by Les Fonds de recherche du Québec (FRQ) and led by Derek Ruths, Professor in the School of Computer Science, aims to advance the efficacy of bot detection in French to mitigate the impacts of online disinformation in Quebec.
Bot detection: A critical tool
Bots are pieces of software that make posts to social media platforms without a human directly involved. Malicious bots typically attempt to appear “human” to other users, making bot detection a critical tool in the fight against disinformation.
“AI-powered bots on social media are often used to spread disinformation. We’d like to remove these bots,” said Ruths. “Unfortunately, they are quite good at pretending to be humans online. This makes them hard to find. As a result, we don’t know how well bot detection methods work and how to make them better.”
According to Ruths, one of the major outcomes of the forthcoming project will be a French-language bot detection dataset, as well as a suite of bots and bot detectors that researchers can use to better understand and improve bot detection methods in French.
Diverse teams
To accomplish this, Ruths looks to XPRIZE for inspiration. XPRIZE is a non-profit organization that hosts public competitions to foster technological advancement, posing big technological challenges and encouraging a wide variety of submissions. Ruths’ project will engage diverse teams of developers, students, scientists, artists, and professionals from across Canada in such a competition, with prizes being awarded to the teams that submit the best bots and bot detectors.
While the project is still in the early planning stages, Ruths said that he aims to launch the competition in the second half of 2024. The work produced from the competition, Ruths said, will play a key role in not only improving bot detection methods in French, but also in advancing researchers’ understanding of bots more broadly.
“This will allow researchers, for the first time, to observe how cutting-edge bots work, how they spread disinformation, and how they can be more effectively detected and managed,” Ruths said.
To learn more or get involved, contact Professor Derek Ruths at derek.ruths@mcgill.ca