As online misinformation grows harder to detect, new artificial-intelligence tools promise to help us separate fact from fiction. But do they actually work?
Not really, according to Dorsaf Sallami. For her doctoral research at Universitรฉ de Montrรฉalโs Department of Computer Science and Operations Research, she examined the limitations of AI systems designed to detect fake news.
Her conclusion: these tools have significant flaws that their technical performance often masks.
…
โCurrent AI systems for detecting fake news are built on a fundamental misconception,โ Sallami said. โWhen AI flags content as false, it doesnโt fact-check as a journalist would. It calculates probabilities based on its training data.โ
In other words, these systems donโt check the facts against reality. They only reflect what theyโve been shown, like a mirror, complete with all the biases and gaps in their training data.
Sallami points to another critical issue: the lack of consensus over what constitutes misinformation.
Share this article
โTo train a system to distinguish fact from fabrication, you have to feed it thousands of examples labeled true or false,โ she explained. โFor simple tasks, like telling a cat from a dog, the labels arenโt controversial. But when it comes to fake news, even experts disagree.โ
Sallami calls this the โground truth problem.โ
