Legislative and Institutional Framework

The Act LXXV of 2025 on the implementation in Hungary of the European Union’s Regulation on AI came into force on 31 November 2025. The purpose of this Act is to develop and implement a national AI strategy and policy, as well as to deepen dialogue and cooperation between stakeholders in the domestic AI ecosystem regarding the development and application of AI, its impact on society, and its regulation. In this Act, the Hungarian legislator created the Hungarian AI Council (AI Council), which (through proposals submitted to the Government via the Ministry of Economy) supports the development and implementation of a national AI strategy and policy regarding the application of AI, the strengthening of technological competitiveness, the widespread and secure use of the technology, and the enforcement of national sovereignty. The Council’s tasks cover, among others, that in the application, development, and regulation of AI, fundamental rights, human autonomy, and the protection of the rights of vulnerable groups – particularly children – are respected. It also coordinates research, development, and innovation activities related to AI in Hungary. Furthermore, the Council develops and continuously reviews the national AI strategy by continuously monitoring the impact of technology on society and the economy, and implements independent professional quality assurance for these strategies.

The Hungarian AI Council was established on 15 December 2025. According to the publicly available statement of the Council, it is committed to the objectives and principles of Regulation (EU) 2024/1689 of the European Parliament and of the Council, the “AI Act”, with particular regard to risk-based regulation, the requirements for high-risk AI systems, and the security of the single market. At the same time, the Council emphasises the importance of striking a balance, noting that the provisions must not undermine the competitiveness objectives of the European Union and its member states, and the Budapest Declaration on the European Competitiveness Pact provides an important reference point in this regard.

The Hungarian AI Council: Directions and Priorities

In its statement, the Council emphasised the importance of ethical standards in connection with the use of AI. In this regard, the Council stated that the development and application of AI will serve the interests of society only if it is implemented in a human-centred, transparent, and accountable manner. The Council emphasised that AI systems must respect human dignity, fundamental rights, and the principle of non-discrimination, in accordance with the values of the Charter and EU law. The Council is initiating the development of a Hungarian AI Code of Ethics and Social Guidelines, based on the principles of transparency, responsibility, accountability, reliability, security, and data protection. The Council undertook to create this document within the first year after its establishment. As of May 2026, this document has not yet been created.

The Council considers it important to continuously examine ethical issues related to AI, including the impact of automation on the labour market, the protection of children, the protection of vulnerable groups, and the preservation of human oversight and control.

The Council also set out the directions of Hungarian national strategy in connection with AI. The Council named three main directions:

• the strategic development goals;
• the development of life skills, education, and social readiness in connection with AI; and
• the creation of an innovation-friendly environment.

As strategic development goals, the Council set out the necessity of the development, capacity expansion, and protection of data centres, high-performance computing (HPC) systems, and cloud infrastructure. The Council considers it important to support Hungarian language and cultural AI models, databases, and language technology developments that ensure the presence of Hungarian language and cultural diversity within the global AI ecosystem. It emphasises the strategic importance of robotics development, with a particular focus on industrial robotics, autonomous systems, logistics and healthcare robots, as well as the paradigm-shifting implications of human-AI-robot collaboration.

In connection with the goal to create an innovation-friendly environment to stimulate competition, the Council is promoting the simplification of regulatory and administrative procedures, the reduction of barriers to entry into the AI market, and the promotion of industry-wide cooperation and investment.

As part of the educational goals, the Council, taking into account the legal provisions regarding the review of higher education institutional regulations and curricula – which support the application of AI – recommends that higher education institutions, building on the sharing of best practices, integrate AI knowledge into their programs in a discipline-specific manner through curriculum additions to be implemented in the coming academic year.

Building the Infrastructure: Hungary in the European AI Ecosystem

In 2025, Hungary joined HunAIFA, a national consortium led by HUN-REN SZTAKI (Institute for Computer Science and Control), which secured designation as an “AI Factory Antenna” under the EU’s EuroHPC programme (European High Performance Computing Joint Undertaking). The Antenna status means Hungary now serves as an official national access point to the JUPITER supercomputer - Europe’s most powerful supercomputer capable of a billion billion calculations per second – with approximately EUR10 million in associated funding available to Hungarian researchers, companies, and public institutions. With limited access to high-performance computing being a barrier to AI development, this is a meaningful step forward.

The development fits well within the priorities set by the AI Council, which has identified data centre capacity, HPC systems, and cloud infrastructure as core national goals. The Council has also stressed the need for Hungarian-language AI models and cultural AI development – recognising that infrastructure investment alone is not enough without corresponding resources that reflect Hungary’s language and cultural context.

Financial Sector: The MNB’s Binding AI Framework

The Hungarian National Bank (MNB) has been one of the most proactive AI regulators in Hungary. In December 2025, it issued Recommendation 13/2025 (XII.3.) on the use of AI by financial organisations – replacing its earlier 2021 digital transformation guidance – setting out what it expects from supervised institutions in terms of AI governance. This is more than soft guidance. Under Act LXXV of 2025, the National Bank was designated as the sectoral AI market surveillance authority for financial services. All credit institutions, insurance companies, investment firms, and other supervised entities must submit an AI sub-strategy to the National Bank by 30 June 2026; this has applied since 1 April 2026.

The sub-strategy must set out concrete milestones and measurable indicators for the safe and responsible use of AI across seven defined areas:

• organisational governance (including the appointment of a designated AI officer with adequate authority);
• data governance;
• model development and procurement;
• risk management;
• cybersecurity;
• staff AI competency; and
• customer communication.

On transparency, institutions are required to inform customers whenever an AI system is involved in a decision affecting them – and must offer a human review mechanism for AI-assisted decisions. There is particular focus on clarity in credit scoring, fraud detection, and anti-money laundering – areas where opaque algorithms raise concerns for regulators and affected individuals alike.

The backdrop for this strategy was a wide-ranging professional consultation conducted by the National Bank, involving – among others – the Hungarian Banking Association (Bankszövetség), the Hungarian Fintech Association (Magyar Fintech Szövetség), and the Digital Companies Alliance (Digitális Vállalkozások Szövetsége), as well as representatives of the National Authority for Data Protection and Freedom of Information (NAIH) and other market participants.

Healthcare Data and AI

Hungary’s National eHealth Infrastructure (EESZT: Elektronikus Egészségügyi Szolgáltatási Tér) has accumulated a substantial health data asset since its launch in 2017: more than ten million citizens’ records, spanning billions of documents including ambulatory records, discharge summaries, diagnostic results, and electronic prescriptions, with millions of transactions processed daily. From 2026, this dataset – anonymised, with the permission of the competent ethics committee – became accessible under Hungarian law for AI development purposes.

The practical potential of the framework is already being explored. In February 2026, researchers at Semmelweis University and HUN-REN announced the pilot launch of iEESZT – a system designed to combine NEAK (National Health Insurance Fund) financing data with EESZT clinical records to construct complete patient pathways and support clinical risk prediction. The stated goal is to make a patient’s full medical history accessible to any treating clinician in a single view, and to flag risks such as missed check-ups or medication interactions.

Intellectual Property at the EU’s Frontier: A Hungarian Referral

Hungary produced what appears to be the first referral to the Court of Justice of the European Union from any EU member state raising AI copyright questions directly. On 3 April 2025, the Budapest Környéki Törvényszék (Budapest Surroundings Regional Court) referred questions to the CJEU in Like Company Kft. v Google Ireland Limited.

The case concerns whether AI-generated content – produced using a copyright-protected dataset without authorisation – infringes the original rights holders’ copyright, and what remedies are available. The referring court concluded that existing EU law did not give a clear enough answer to decide the case itself.

A hearing before the CJEU took place on 10 March 2026. An Advocate General Opinion is expected in September 2026, with a judgement likely in late 2026 or early 2027. The case is closely watched across the EU: the AI Act deliberately left copyright questions aside, and there is currently no harmonised EU rule on whether rights holders can prevent their works from being used to train AI systems. Whatever the CJEU decides will have direct relevance for IP practitioners and AI developers throughout Europe.

Hungary’s Intellectual Property Office (SZTNH: Szellemi Tulajdon Nemzeti Hivatala), which sits on the AI Council, has been tracking AI and IP policy at the EU level. The Like Company case has placed Hungary at the centre of a legal question with consequences well beyond its borders.

AI and the Legal Profession

The Hungarian legal profession is adapting to AI at a time when both regulation and professional guidance are still catching up with practice.

In March 2026, the Ministry of Justice launched jog.gov.hu, an AI-powered legal information platform. This website handles natural language queries about Hungarian law and returns structured answers from official sources. The platform is open to the public rather than restricted to lawyers, but its launch signals that the government sees AI as a legitimate interface for accessing legal information.

At the European level, the Council of Bars and Law Societies of Europe (CCBE) published guidance on the use of generative AI by lawyers in September 2025. A Hungarian translation appeared on the website of the Hungarian Bar Association (Magyar Ügyvédi Kamara, MÜK) in March 2026. As of May 2026, MÜK has not issued its own position on AI use by Hungarian lawyers, leaving the CCBE document as the main reference point available through official professional channels. The absence of domestic bar-level guidance may lead to uncertainty, though, particularly around the national level interpretation of confidentiality, the duty of competence, and oversight of AI-assisted outputs when it comes to the services of legal professionals.

In practice, Hungarian law firms are divided. Some have already integrated AI into document review, contract drafting, and legal research; others are waiting for clearer professional guidance before committing. This split mirrors what is happening in legal markets across the EU. More and more, AI services are becoming available for legal professionals specifically in Hungary. Many of them run on top of Open AI’s ChatGPT. Some are restricted to searching in legislation databases and judicial decision databases, others work on more comprehensive legal tasks such as drafting. Each requires careful supervision, underpinned by the knowledge and experience of a practising lawyer, and while these services may help and shorten the preparation time for a given legal document, they are far away from replacing human work and intelligence when it comes to reasoning.

Challenges Ahead

Despite rapid progress, some gaps remain in Hungary’s AI governance landscape. The AI Code of Ethics and Social Guidelines that the Council committed to producing in its first year has not yet been published.

With the Council established in December 2025, less than a year ago, this is an important gap-filling development to look forward to, given how centrally ethical standards featured in the founding statement.

The regulatory sandbox under Act LXXV of 2025 is also not yet operational. The provisions governing it – Section 10 of the Act – do not enter into force until 2 August 2026. Until then, companies seeking to test AI products under reduced regulatory requirements have no formal route to do so.

Implementation of the National AI Strategy for 2025–2030 remains uneven across government agencies. The financial sector, where the National Bank has moved quickly to issue binding requirements, is more of an exception than the rule. Agency-level AI sub-strategies required under Act LXXV of 2025 are still being developed in most sectors.

A significant structural variable entered the picture in April 2026, when Hungary’s parliamentary elections resulted in a change of government. The young Tisza party took office with an ambitious digital agenda: its election programme envisaged a personal AI assistant for every Hungarian citizen, a full Hungarian-language AI model, and cybersecurity audits across all state institutions. What the new government has not yet set out is how it intends to handle the AI governance architecture it has inherited: the composition and mandate of the AI Council, the allocation of AI supervisory responsibilities across the new 16-ministry structure, and the activation of the regulatory sandbox. These questions remain open as of the date of writing.

Summary

Hungary’s AI landscape has changed considerably over the past year. The legislative framework is in place, an AI Council has been created, and sectoral regulators – particularly in the financial sector – have started turning new rules into concrete requirements. Infrastructure is developing, and there is also active judicial contribution to the clarification of EU law, with a Hungarian court having sent a case to the CJEU that may shape EU AI copyright law for years.

Some important pieces are still missing, though. An AI ethics code is not yet published, the regulatory sandbox is not yet running, and lawyers – like professionals in many fields – are navigating AI adoption without much formal guidance from their professional bodies.

The next year will surely bring important developments and fill some gaps. The sandbox is expected to be activated, the ethics code to be published, and the Advocate General’s Opinion, as well as the CJEU’s judgement in Like Company v Google, will all shape how Hungary’s AI governance framework develops in practice – and how much of its early ambition translates into workable rules.