Студентка из США создала ИИ для помощи активистам в борьбе с авторитарными режимами

On co-op in Serbia, she helped build an AI to give activists worldwide a leg up Ayla DiBattista was searching for a co-op opportunity that would combine her passion for international affairs, travel and social justice. A globe-trotting sophomore with family scattered across Europe, DiBattista was itching to return overseas and immerse herself in a foreign culture: “I always kind of wanted to live on my own and see if I could do it,” she said. Of the more than 20 countries she’s visited, her favorite destinations include Montenegro, Sweden, Hungary and Portugal. So when, in 2025 and fresh off a summer Middle East Dialogue of Civilizations in Egypt, Cyprus and the United Arab Emirates, political science professor Denis Sullivan suggested that she look to a little-known nonprofit in the Balkans in southeastern Europe, DiBattista was taken with the idea. That organization, the Center for Applied Nonviolent Actions and Strategies, or CANVAS, has equipped activists around the world with the tools to resist oppressive governments through nonviolent means. Founded in 2004, it traces its roots to Otpor!, the student movement that helped topple former Serbian President Slobodan Milošević, who was the first sitting head of state indicted for war crimes. CANVAS has since grown into a “global school for activists,” offering workshops, consultations and training to civil society actors worldwide, according to the organization’s website. DiBattista got to work with Srdja Popović, one of the architects of Serbia’s nonviolent revolution. Popović, co-founder of CANVAS, formerly served the Serbian National Assembly and was a 2012 Nobel Peace Prize nominee. DiBattista worked under Popović’s supervision and in coordination with computer scientists building a platform that contained academic literature on the history of nonviolent movements. “I really appreciate working with him,” DiBattista said. “It was really cool seeing our project come to life.” The goal was to give Gene, the proprietary AI model named after American political scientist and nonviolent action expert Gene Sharp, a comprehensive understanding of how such movements operate across different political environments. Sharp founded the nonprofit Albert Einstein Institution in 1983, where many activists associated with Otpor! were trained. CANVAS’ work on Gene is one of the few known efforts on the part of a nonprofit to harness the power of AI to combat state repression and empower local resistance movements in societies around the world. Authoritarian governments such as China, Russia and Iran have leveraged artificial intelligence for surveillance, censorship and information control; and researchers and pro-democracy advocates are beginning to explore whether the same technology can help buoy their own local efforts to defend democratic principles and resist oppression. The project work was done at CANVAS headquarters in Belgrade, where DiBattista said her primary task was “research gathering” across different subject areas: coalition building, mobilization and outreach. Using her Northeastern Claude subscription, DiBattista spent a good chunk of her days compressing dense texts into structured training documents, helping equip the in-house AI to retrieve practical guidance for activists anywhere “who are looking to start a movement,” she said. Training AI tools using AI tools — an increasingly common aspect of workflow today — still depends on human beings to carefully fact-check the output and ensure the final product aligns with a larger vision, she said. Now back in the United States, DiBattista is plotting her next move. She’ll be entering her third year in the fall; and though she had only minimal experience with AI systems before CANVAS, she’s become something of an AI specialist. In June, DiBattista started working at the Philadelphia-based Welcoming Center, a nonprofit that helps immigrants access economic opportunities and integrate into their communities. She is helping the newcomers navigate everything from paperwork and public benefits to employment and language barriers — all hurdles familiar to those recently arriving in the country. Predictably, she went to work figuring out ways of using artificial intelligence to make those systems more accessible to immigrants. Though she is still ambivalent about AI’s ethical impact, DiBattista said the technology should be available to everyone. “I think it’s very important for all kinds of people to be able to use it, especially those in primarily disadvantaged positions,” she said, “whether that’s under a repressive regime or immigrants looking to find a home in a new country.”