Oren Tsur, an expert on artificial intelligence, is more than just a little excited these days.
“Really, every research or project I’m involved with currently I think is super exciting,” the Israeli researcher said. “And I could talk about it for hours.”
Tsur is an expert on using AI to analyze — and hopefully mitigate — hate speech and misinformation online. He’s in the Bay Area this week where he will participate in a panel on AI at the Oshman Family JCC in Palo Alto.
The Feb. 23 event, open to the public, is co-sponsored by Americans for Ben-Gurion University and the Jewish High Tech Community, a networking organization in Silicon Valley. Nir Peled, a general manager in the field of artificial intelligence at Intel, will moderate the panel, which will also feature Eric Amram, CEO of the event-planning software company Evenium.
Tsur will discuss his research and what can be done to make the internet a safer space for open, constructive communication.
“We don’t want people to agree, we just want to maybe stop polarization,” he said.
Tsur’s research uses computer algorithms to understand how misinformation, propaganda and hate speech spread on the internet, whether planted deliberately — such as with Russian misinformation about Ukraine — or evolving naturally as extremists find each other and organize online.
The process is complicated. Tsur explained that first experts in linguistics, education and sociology annotate text, marking it so that an AI can be trained on the material to analyze what makes a message hate speech or misinformation. That allows researchers to look for patterns.
“We have a number of examples, like hundreds or thousands of examples, that were annotated by humans, by trained, actual annotators,” Tsur said. “And we use cutting-edge AI to do it on a large scale.”
Tsur’s home university, Ben-Gurion, has recently launched an interdisciplinary center where sociologists, linguists and AI experts can work together to understand antisemitism, hate speech, bias and misinformation. Tsur is the center’s chair.
“Maybe the most significant avenue of research that’s going to happen there is about election interference and how states, nation states, are weaponizing social media,” he said. “It could be for election interference or for just disinformation propaganda. We see it a lot in the war in Ukraine now.”
Tsur considers earlier election interference campaigns, such as one that China ran in Canada in 2019, trial runs. And he thinks it’s going to get worse.
With a number of elections coming up in the next year in Europe, “we suspect that it will be contentious, especially because of what’s happening in Ukraine and the Baltic, the Nordic countries and the debate over NATO and about natural gas and energy,” he said.
Tsur acknowledges the challenges of keeping up with the rapid speed of AI developments.
“There are so many technological changes in recent years that provide attackers more power to do things more efficiently or in a more covert way,” he said. “But also for us, we have stronger tools to try and detect and mitigate. So it’s a cat-and-mouse thing. I don’t think it’s going to stop.”
“The Power of AI” will run from 6:30 to 9:30 p.m. on Feb. 23 at the Oshman Family JCC’s Freidenrich Hall, 4th floor, 3921 Fabian Way, Palo Alto. $20 (includes dinner). For more information, click here.