Introduction
The European Commission has released draft guidelines to aid developers in adhering to the AI Act, a regulation that governs artificial intelligence usage within the European Union. These guidelines aim to provide legal clarity and promote consistent application of the Act, particularly concerning AI systems that pose unacceptable risks to fundamental rights.
Description
The European Commission has published draft guidelines to assist developers in complying with the AI Act (Regulation (EU) 2024/1689), which governs the use of artificial intelligence within the European Union as the market expands. This comprehensive 135-page document aims to provide legal clarity for businesses regarding AI systems deemed to pose “unacceptable risks” to fundamental rights [2]. Although non-binding [2], the guidelines promote a consistent and effective application of the Act by offering legal explanations and practical examples of unacceptable AI practices [1]. The provisions of the AI Act regarding AI literacy and the prohibition of certain AI systems became effective on February 2, 2025, marking the beginning of compliance requirements [6]. It categorizes AI systems into various risk levels [5], including prohibited [5] [8], high-risk [2] [3] [5] [6], and those with transparency obligations [5], with the goal of fostering innovation while safeguarding health, safety [2] [3] [5], and fundamental rights [2] [3] [5] [8].
The guidelines clarify elements of the AI Act [9], particularly those prohibiting risky applications of AI. They identify eight specific practices deemed unacceptable due to their potential risks to safety, European values [5], and fundamental rights [2] [3] [5] [8]. Among the prohibited practices are manipulative techniques, exploitation of vulnerabilities based on age or economic status [3], social scoring [1] [2] [3] [5] [8], individual predictive policing based solely on profiling [3], untargeted scraping of facial images for facial recognition databases [3], emotion recognition in workplaces and schools [3], biometric categorization using sensitive characteristics [3], and real-time remote biometric identification in public spaces by law enforcement [3]. Notably, while the use of live facial recognition is banned, exceptions exist for law enforcement and border control in serious crime investigations, such as terrorism, provided that police obtain authorization and do not deploy these technologies in a broad [2], untargeted manner [2]. Additionally, the guidelines explicitly bar the use of AI to infer employee emotions or assess individuals’ risks of criminal behavior, as well as AI-enabled “dark patterns” that manipulate individuals into actions they would not otherwise take [8], including chatbots employing subliminal messaging to influence harmful financial decisions [8].
Violations of the prohibitions outlined in Article 5 of the AI Act can result in significant penalties [7], including fines of up to €35 million or 7% of the previous year’s global turnover per violation [7], whichever is higher [2] [4] [8]. Enforcement of these rules will be the responsibility of National Market Supervisory Authorities and the Data Protection Authority [3], which must be designated by August 8, 2025 [3]. Until that date, public enforcement will not be in effect [3].
To mitigate risks associated with prohibited AI practices [7], companies should proactively identify [7], assess [7] [8], and document the AI systems utilized in their operations [7], particularly those that may be categorized as prohibited [7]. Given the ambiguous nature of the prohibited activities described in the AI Act [7], it is crucial for companies to maintain clear documentation justifying their determination that specific AI uses do not fall under the prohibitions [7].
The guidelines reflect the EU’s commitment to a safe and ethical AI environment and are pending formal adoption. Subsequent obligations under the Act will be phased in over time, with most requirements for high-risk AI systems and specific transparency rules becoming applicable on August 2, 2026 [6], and obligations related to high-risk systems outlined in Annex I taking effect on August 2, 2027 [6]. Some AI systems and models already available may qualify for exemptions or extended compliance timelines [6]. Authoritative interpretations of the guidelines will ultimately be determined by the Court of Justice of the European Union, ensuring clarity and consistency in the application of the AI Act across the EU.
Conclusion
The draft guidelines by the European Commission underscore the EU’s dedication to fostering a secure and ethical AI landscape. By providing clarity and promoting consistent application of the AI Act, these guidelines aim to balance innovation with the protection of fundamental rights. The phased implementation of the Act’s requirements and the potential for significant penalties highlight the importance of compliance for businesses operating within the EU. Ultimately, the Court of Justice of the European Union will play a crucial role in ensuring uniform interpretation and application of these guidelines across member states.
References
[1] https://www.lexisnexis.co.uk/legal/news/commission-issues-guidelines-on-prohibited-ai-practices-under-ai-act
[2] https://www.biometricupdate.com/202502/eu-issues-guidelines-clarifying-banned-ai-uses
[3] https://www.ansa.it/english/news/sciencetecnology/2025/02/05/ai-act-bans-come-into-force-eu-guidelines-for-firms-states78cba6d0-a0e4-426a-a006-eb8b0871b48d.html
[4] https://techcrunch.com/2025/02/04/eu-puts-out-guidance-on-uses-of-ai-that-are-banned-under-its-ai-act/
[5] https://digital-strategy.ec.europa.eu/en/library/commission-publishes-guidelines-prohibited-artificial-intelligence-ai-practices-defined-ai-act
[6] https://natlawreview.com/article/rules-ai-literacy-and-prohibited-systems-under-eu-ai-act-become-applicable
[7] https://www.lw.com/en/insights/upcoming-eu-ai-act-obligations-mandatory-training-and-prohibited-practices
[8] https://www.siliconrepublic.com/machines/ai-act-eu-guidelines-prohibition
[9] https://news.bloomberglaw.com/artificial-intelligence/eu-releases-guidance-clarifying-aspects-of-its-landmark-ai-act
