AI-generated disinformation and deepfakes were highlighted as a growing challenge for democratic debate and elections during an Interparliamentary Conference on Artificial Intelligence within the framework of the Cyprus Presidency of the Council of the EU, hosted by the House of Representatives in Nicosia.
Addressing the 15 January conference, the EU Commissioner for Democracy, Justice, the Rule of Law and Consumer Protection Michael McGrath set out how the EU’s digital framework is intended to address risks linked to artificial intelligence.
McGrath thanked House President Annita Demetriou and the organisers for convening the discussion during the early period of the Cypriot Presidency.
He referred to areas where AI systems can support consumers and public administration, including comparing prices, identifying scams, assisting courts with document analysis and case management, and helping to detect disinformation and foreign information manipulation. At the same time, he argued that public policy needs to account for situations where AI is used in ways that can undermine democratic processes and public trust.
Pointing to concerns about false information, McGrath said the information environment faces increased pressure from attempts to spread disinformation, manipulate public debate and interfere in elections, with generative AI enabling new techniques. He referred to “hyper-realistic deepfakes” and automated, personalised disinformation campaigns as developments that can be deployed quickly and at scale.
EU’s approach: Setting “guardrails” through a package of legislation and measures
McGrath said the EU’s approach is to set “guardrails” through a package of legislation and measures aimed at platform accountability and user protection. He listed the main elements of the EU’s digital framework as: the General Data Protection Regulation (GDPR), the Digital Services Act (DSA), the Code of Conduct on Disinformation, the Political Advertising Regulation and the European Media Freedom Act.
In outlining impacts for citizens, he said the GDPR is intended to strengthen control over personal data, while the DSA sets obligations for online platforms in areas such as illegal content and risk mitigation, alongside protections for freedom of expression. He added that the AI Act is intended to set requirements so that AI systems are developed and used in ways aligned with fundamental rights, and that the European Media Freedom Act is aimed at supporting access to reliable information and protecting independent journalism. McGrath said the AI Act prohibits certain practices, including AI-enabled manipulative or deceptive techniques and systems designed to exploit vulnerabilities.
He also pointed to obligations to enable detection and labelling when content has been artificially generated or manipulated, including labelling deepfakes and certain AI-generated publications on matters of public interest and noted that the European Commission is working on guidelines on classifying high-risk AI systems, including those linked to electoral processes.
McGrath said the DSA places legal obligations on very large online platforms to address coordinated inauthentic behaviour and to mitigate disinformation risks linked to recommender systems. He also referred to the Code of Conduct on Disinformation as a framework bringing together industry and civil society, including work on labelling AI-generated and AI-manipulated content, increasing transparency in recommender systems, and limiting financial incentives for disinformation.
McGrath said the Political Advertising Regulation sets EU-wide transparency and due-diligence standards, intended to help people identify when political advertising is paid for, who paid for it, how much was paid, and whether ads are targeted to specific audiences. He also referenced restrictions on ads from non-EU sponsors during the three months before an election or referendum in an EU Member State, describing this as a measure aimed at reducing risks of foreign information manipulation and interference.
European Democracy Shield and cooperation with the Cypriot Presidency
McGrath said the European Democracy Shield, presented in November, is organized around priorities that include safeguarding the integrity of the information space, supporting democratic institutions and elections, and strengthening societal resilience and citizen engagement.
He highlighted plans for a European Centre for Democratic Resilience to pool expertise, share information and improve coordination when disinformation or foreign information manipulation arises, adding that work is underway to launch the center in cooperation with the Cypriot Presidency.
He also referred to cooperation through the European Cooperation Network for Elections and forthcoming guidance on the responsible use of AI in elections.
McGrath said the Commission plans to propose a Digital Fairness Act later this year, focused on consumer protection in the online environment, including misleading practices such as “dark patterns”, addictive designs and deceptive marketing, with attention to protecting children online.
(Source: CNA)
