Artificial intelligence could not only produce deepfakes and political disinformation but also influence voters’ decisions at the ballot box. According to experts, algorithms rather than political parties may determine the outcome of the upcoming elections in Poland.
Jakub Szymik, founder and president of the Digital Democracy Observatory Foundation, warns that AI models can adapt to individual voters, conversations, and user needs. – This technology can influence elections and tip the scales in one direction or the other – he stated in an interview with WNP.PL.
The warning comes as Poland still lacks precise rules for labeling AI-generated content, political influencer advertising, and the mass use of artificial intelligence in campaigns. A bill prepared in the Senate would allow imposing fines on online platforms of up to 6 percent of their annual global revenue.
AI models and the risk of political bias
Szymik pointed to an incident involving Grok, a chatbot operating on X, which explicitly declared it would vote for the Confederation party. – Grok clearly stated that it would vote for Konfederacja – he recalled. This is particularly notable given that Meta removed the party’s Facebook page in 2022.
Concerns extend beyond individual chatbots. During the UN Global Dialogue on AI governance, a strong warning was issued that AI could disrupt democratic processes. Reports from the UK elections showed models hallucinating about election dates, political programs, and even the existence of candidates. – This is very difficult to monitor – Szymik admitted.
During the European Parliament elections, thousands of deepfake posts appeared showing Ursula von der Leyen or Stalin dancing on the grave of the European Union. – No individual citizen would be able to generate such a quantity of content – he emphasized.
Poland’s legislative response
The Senate has initiated work on an act implementing EU regulations on political advertising. If signed by the president, the law would give the National Electoral Commission the power to impose significant penalties on digital platforms. However, there is a risk of a veto, similar to the one that blocked the Digital Services Act implementation.
Digital platforms like Meta, Google, TikTok, and X currently have varying approaches to political advertising. Their declarations often differ from actual practices. The regulations would require labeling political advertising, disclosing who is behind it, and prohibit targeting based on special categories of personal data.
Six institutions would be responsible for enforcement, including the National Electoral Commission, the Personal Data Protection Office, and the Office of Competition and Consumer Protection. – It will be effective if these bodies want to deal with it. So far, we have the impression that they fear competency competition – Szymik noted.
Źródło: WNP.PL, Fot. dee karen/Shutterstock, Fundacja Obserwatorium Demokracji Cyfrowej
