- 2024
- 2025
- 2026
- 2027
Entry into force of the AI Act
After a three-year legislative procedure, the EU’s Artificial Intelligence Act (AI Act), a set of new rules to regulate Artificial Intelligence, was published in the Official Journal and entered into force on 1 August 2024. This marks the start of the implementation period, with some exceptions, with all rules becoming fully applicable only in 2026.
Application of rules on prohibited AI systems
The first batch of rules under the AI Act that become applicable are the prohibitions of certain AI systems, including those used for subliminal manipulation and for categorisation according to political opinions. Guidelines have also been published to clarify the prohibitions with examples of what should and shouldn’t be included under the banned applications.
Finalisation of the Code of Conduct on General Purpose AI (GPAI) models
Under the Artificial Intelligence Act, a voluntary Code of Conduct has been shaped to clarify and complement the rules on General Purpose AI. The work within the General Purpose AI Working Group took place from autumn 2024 until spring 2025.
The European Commission puts forward a proposal on a Digital Omnibus
On 19 November 2025, the European Commission published a proposal for a regulatory simplification package on a Digital Omnibus. The new proposal aims at simplifying existing rules to boost innovation for EU businesses, in particular with modifications in the area of data protection (General Data Protection Regulation) and Artificial Intelligence (AI Act). However, the proposal might have the reverse effect and create more legal uncertainty and difficulties for companies to comply with the new rules. It is also a dangerous precedent as it proposes to scale back existing protections of fundamental rights in the EU, for example, on personal data. The proposal still needs to be negotiated and amended by the European Parliament and the Council of the EU in the next few years.
Publication of draft guidelines on high-risk AI systems by the European Commission
The rules for high-risk AI systems are at the core of the Artificial Intelligence Act, with applications such as AI systems used to influence elections undergoing risk assessments before being put on the market. To clarify and complement such rules, the European Commission issued guidelines in May 2026.
Full application of all AI Act rules (except the ones for high-risk AI applications)
The EU’s Artificial Intelligence Act becomes fully applicable as of October 2025, with the exception of the rules for high-risk AI applications that will only apply in 2027.
Entry into application of the rules on high-risk AI systems
The rules for high-risk AI systems are at the core of the Artificial Intelligence Act, with applications such as AI systems used to influence elections to undergo risk assessments before being put on the market. They also include AI systems intended to be used to influence elections as high-risk. The entry into application of these rules, according to the AI Omnibus, will be December 2027 for standalone high-risk applications and August 2028 for AI embedded in other products.
Artificial Intelligence
Artificial Intelligence has been at the top of the EU agenda since the start of the 2019-2024 mandate, with the proposal for an Artificial Intelligence Act (AI Act). The discussions around it, however, intensified further in 2023, with the boom of generative AI systems. From deep fakes of politicians circulated online to AI-generated disinformation campaigns, from AI systems used for microtargeting of political ads to chatbot-based AI systems to provide voter assistance, AI applications with the potential to impact democracy and elections are countless. EPD is therefore working on assessing the effectiveness of mechanisms in the AI Act to mitigate the risks posed by such applications.
The Digital Services Act
The Digital Services Act (DSA) is an EU horizontal framework of rules put forward to create a safer online space. As many challenges posed by online platforms and regulated via the DSA can have an impact on democratic civic discourse and electoral processes, such as the spread of disinformation, coordinated attacks against civil society or the lack of transparency on how recommender systems work, having a DSA that works for democracy is of utmost importance for EPD.
Transparency and targeting of political advertising
Political advertising online has been historically the main focus of EPD’s work in the digital sphere. Since the Cambridge Analytica scandal, it has become increasingly evident how online practices such as micro-targeting of political ads can steer public opinion ahead of elections. When the European Commission put forward its proposal for a Regulation on Transparency and Targeting of Political Advertising (TTPA), EPD took an active role in leading the advocacy work on this file by promoting more transparency and accountability for both sponsors and online platforms.
Our work on these files is funded by Civitates and the Open Society Foundations.