Artificial intelligence is fundamentally reshaping the way information is created, distributed, and verified. At the same time, the very technology that facilitates the spread of disinformation also provides new tools for identifying and countering it. This was one of the key conclusions of the third panel of the ICDC Conference, “Moldova’s response to disinformation: lessons and experiences”, titled “Technology, artificial intelligence and digital tools in support of truth and information integrity.”

Participants emphasized that, despite the rapid development of artificial intelligence tools,  human judgment remains essential in verifying information and making decisions. “These tools are very useful, but it is the human being who must make the final decision,” said WatchDog expert Andrei Curăraru, noting that artificial intelligence should be regarded as a support tool rather than a replacement for human expertise.

The experience of StopFals.md has shown that technology has already become indispensable to fact-checking. According to Editor-in-Chief Mihai Avasiloaie, AI-powered tools make it possible to monitor in real time the narratives promoted by actors spreading disinformation and help detect manipulative content more quickly. However, the speed at which false information circulates remains a major challenge. As he noted, fact-checking continues to be a rigorous process and cannot be made less thorough simply because disinformation spreads so rapidly.

The speakers also pointed out that the greatest challenge is not obvious falsehoods, but narratives designed to exploit people’s emotions. “I believe the most difficult messages to debunk are emotional ones,” Mihai Avasiloaie said, explaining that such messages cannot be effectively countered merely by presenting facts or figures, but require a deeper understanding of how they are constructed and disseminated.

Victoria Olari of Atlantic Council Moldova highlighted that research on disinformation is evolving rapidly, with the focus shifting from the analysis of individual content to the analysis of networks and coordinated behavior. She also warned about an emerging risk known as “data poisoning”, whereby datasets used to train AI systems can be contaminated with manipulative information, potentially influencing the responses generated by chatbots.

The winners of the “Deepfake” hackathon, organized in April in Cahul, presented a prototype application that automatically analyzes news articles, images, and videos and provides users with explanations of suspicious elements identified by its algorithms. The project demonstrates that locally developed technological solutions can support the work of journalists and civil society organizations.

The discussion concluded on an optimistic note. Technology and artificial intelligence should not be viewed as threats, but as opportunities that, when used responsibly and ethically, can strengthen society’s resilience against disinformation. Participants stressed that the success of this effort depends on collaboration among experts, journalists, developers, civil society, and international partners, as well as continued investment in digital skills and innovation.