Researcher Mateusz Łabuz, from the IFSH, recently shared insights in a Help Net Security interview regarding the delicate balance between utilizing AI for personalized political campaigns and safeguarding voter privacy. Discussing the potential of AI in fact-checking, the regulatory landscape, and the impact of AI on campaign strategies in authoritarian regimes, Łabuz shed light on various aspects of AI’s role in the political sphere.
When asked about how campaigns can navigate the use of AI for personalization while considering voter privacy concerns, particularly in regions with weaker data protection laws, Łabuz emphasized the importance of establishing clear regulations. By implementing rules that govern the actions of political entities during campaigns, including guidelines on information collection, a protective barrier can be erected against potential abuses. He highlighted the efforts within the European Union to create legal frameworks that promote informed decision-making among citizens, such as the transparency and targeting regulations for political advertising. Collaboration between regulators and digital platforms was also emphasized to ensure data protection and limited personalization.
Regarding the early detection of AI-driven disinformation in the crucial final days of an election, Łabuz mentioned the existence of early detection systems like content analysis and social networking algorithms. These systems are designed to monitor the spread of disinformation in real-time, with a focus on combatting disinformation at its source—the digital platforms that amplify such content. While some jurisdictions have introduced regulations to tackle disinformation close to elections, effective enforcement is key to their success. He suggested intensifying fact-checking activities and investing in content moderation during critical periods to mitigate the impact of AI-generated disinformation.
In terms of existing regulations like the EU’s Digital Services Act and their effectiveness in curbing AI-generated disinformation, Łabuz acknowledged the challenges of assessing the regulations’ impact so soon after implementation. While there have been positive developments in terms of transparency and reporting obligations for digital platforms, more time and data are needed to evaluate the regulations’ efficacy. He also raised concerns about gaps in the regulatory framework, particularly in navigating the fine line between combating disinformation and upholding principles of freedom of speech and expression.
On the possibility of developing AI-driven tools for real-time fact-checking during campaigns, Łabuz noted ongoing advancements in sentiment analysis and content detection. However, he highlighted barriers to large-scale implementation, including language complexity, access to updated data, algorithm bias, and the potential for errors in automated content moderation. Stressing the importance of human oversight and collaboration with technology, he also mentioned tools for detecting synthetic media and the EU AI Act’s provisions for marking such content at the provider level.
Lastly, when discussing how global AI trends are impacting campaign strategies in authoritarian regimes, Łabuz emphasized AI’s role as a tool for exerting control over citizens. While AI can empower activists to counter authoritarian manipulation, the overarching trend seems to reinforce existing power imbalances and strengthen authoritarian control over political narratives. Despite a few instances where AI has been used creatively to challenge authoritarian rule, the overall trajectory suggests a reinforcement of control mechanisms within authoritarian regimes.