Hungary’s Facial Recognition Proposal Likely Violates the EU AI Act

As EU Member States prepare to implement the EU AI Act, Hungary’s latest legislative proposal has raised significant compliance concerns. A recent amendment to the Hungarian Child Protection Act, reportedly intended to ban pride events and enable the use of facial recognition to identify participants, raises important questions about alignment with the EU’s AI governance framework.

The proposed use of real-time facial recognition in public spaces likely stands in conflict with Article 5 of the EU AI Act, which classifies such applications as prohibited, except under strictly defined scenarios. These include the prevention of imminent terrorist threats or the investigation of severe criminal activities tied to national security. Use cases outside these parameters are not permitted under the regulation and we emphasize, that the rules related to prohibited AI systems are already in effect and applicable for all Member States.

According to Dr Laura Caroli, a key AI Act negotiator for the European Parliament, the proposed deployment would not only exceed the AI Act’s scope for exemptions but would also contravene its foundational principles. Even if a national security justification were claimed, she notes, it would still likely fall short of the legal and procedural thresholds outlined in the Act.

For both public institutions and private enterprises across Europe, this case is a critical reminder: the AI Act is not just a technical framework, it is a legal and ethical safeguard for fundamental rights. The classification of biometric identification in public spaces as prohibited AI is intentional and rooted in protecting individuals from intrusive surveillance and misuse of advanced technologies.

At Novius Consulting, we emphasise that responsible AI begins with rigorous compliance and governance. Deploying powerful tools such as facial recognition demands more than just technological readiness, it requires a clear legal basis, risk assessments, and an unwavering commitment to European values such as privacy, transparency, and non-discrimination.

Hungary’s proposed approach, if enacted, could challenge the consistency of AI governance across the EU and risk setting precedents that erode public trust. It also serves as an urgent call to action: organisations and institutions must proactively assess whether their AI use cases, particularly those involving biometric or surveillance technologies, fall within the legal boundaries of the AI Act, as violations could result in penalties of up to €35 million or 7% of global annual turnover.

As parts of the EU AI Act is already in effect, Novius Consulting supports organisations in preparing robust compliance strategies, and embedding trustworthy practices into AI systems from day one. In a landscape where technology evolves quickly, only strong governance ensures long-term value and legal certainty.