The European Union is introducing updated regulations governing artificial intelligence, which will take effect on August 2, 2026. A notable change is a requirement for transparency and labeling of content created or significantly altered through AI, as stipulated in Article 50 of the regulation. Such content must be clearly labeled, especially when it involves deepfakes depicting real events or featuring politicians or public figures.
However, the rules do not apply to content unless the user has reasons to perceive it as genuine. Exceptions are made for media: materials verified by journalists or editors, under editorial responsibility for socially important topics, do not require labeling. Nonetheless, editorial offices are obliged to implement internal procedures that are audited to confirm human oversight of the process.
Simultaneously, the European Parliament has introduced a prohibition on generating intimate or violent content involving children without the consent of the concerned individuals, referred to as ‘nudification.’ Companies have until December 2, 2026, to adapt their systems to comply with these requirements. The Polish Personal Data Protection Office (UODO) emphasizes that existing legislation prohibits the digital processing of photographs without appropriate consent, underscoring the need to adhere to current norms.
Representatives from the advertising sector have expressed concerns about potential over-regulation, arguing that extensive labeling of AI-created content might diminish consumer attention to such materials.
Violations of these new rules will incur significant penalties, ranging from 7.5 to 35 million Euros or up to 7% of the company’s annual turnover. While the majority of the provisions become mandatory on August 2, 2026, full implementation of the regulation is expected by August 2, 2027. Consequently, European businesses must choose between promptly adapting their operational processes and risking substantial financial losses.