Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49

Notice: Undefined index: en in /home/user_1/web_1/web/app/core/router/RouterTranslates.php on line 49
New Decree on the Conformity Assessment of High-Risk AI Systems - IT Law, Drafting and reviewing contracts,Discussions in regards of Domain names, Counselling on software- and licensing agreements

Legal blog

New Decree on the Conformity Assessment of High-Risk AI Systems

2026-01-06 | IT Law

New Decree on the Conformity Assessment of High-Risk AI Systems

1. The Place of the Decree within the Hungarian AI Regulatory Framework

The decree, which enters into force on 23 January 2026, aims to establish detailed rules concerning the designation of bodies responsible for certifying the prior technical conformity of high-risk AI systems. This regulation is directly linked to the domestic implementation of Regulation (EU) 2024/1689 (the AI Act).

In designing the implementation model, the Hungarian legislator relied on existing conformity assessment structures. Accordingly:

  • as a general rule, the designation of conformity assessment bodies is governed by Act CXXXIII of 2009 on the Activities of Conformity Assessment Bodies;

  • the authority responsible for designation is the AI Notifying Authority, which, in this capacity, builds upon existing accreditation and administrative mechanisms.


2. Scope of Application – To Whom and to What Does the Decree Apply?

The scope of the decree expressly covers:

  • conformity assessment bodies responsible for verifying compliance with requirements applicable to high-risk artificial intelligence systems; and

  • the procedure for the designation of such bodies.

It is important to emphasise that the decree does not independently define what qualifies as a high-risk AI system. This continues to be governed by the AI Act and the related body of EU law. The Hungarian implementing decree focuses exclusively on the procedural and institutional framework.


3. Liability Insurance

One of the key provisions of the decree is that it makes liability insurance for conformity assessment bodies an integral part of the accreditation procedure.

This regulatory approach conveys a clear legal policy message: the conformity assessment of high-risk AI systems is an activity involving substantial liability risks, and the legislator therefore requires the existence of adequate financial safeguards.


4. Fees for the Designation Procedure

The decree lays down detailed rules on the administrative service fees payable for the designation procedure, which in practice represent a non-negligible cost factor for conformity assessment bodies.

The fees are as follows:

  • HUF 250,000 per accredited status; and

  • an additional HUF 20,000 for each product group and each conformity assessment procedure (module).

In the case of data modification, a uniform administrative service fee of HUF 20,000 is payable.

The legislation also precisely defines the method and timing of payment, as well as the earmarked use of the fees—the amounts collected may be used exclusively to cover personnel and material costs related to the designation procedure.


5. What Does This Mean for Companies Developing and Deploying AI Systems?

Although the formal addressees of the decree are conformity assessment bodies, its indirect impact on market participants is also significant.

Developers and distributors of high-risk AI systems must take into account that:

  • conformity assessment is conducted within a strictly regulated and institutionalised procedure;

  • only designated and accredited bodies are entitled to certify prior conformity;

  • the costs and duration of such procedures must be factored into project planning in advance.

Accordingly, the decree indirectly encourages undertakings to incorporate legal and compliance considerations into the design of their AI systems already at the development stage.


Our colleagues continuously monitor developments in the legal regulation of artificial intelligence. If you are an AI developer and have questions regarding the legal implications of AI development, please feel free to contact our law firm.

New Decree on the Conformity Assessment of High-Risk AI Systems

2026-01-06
New Decree on the Conformity Assessment of High-Risk AI Systems

The implementation of European Union regulation concerning artificial intelligence has entered a new phase in Hungary.
Decree No. 44/2025 (XII. 23.) of the Ministry for National Economy (NGM) constitutes an implementing regulation related to high-risk artificial intelligence systems, setting out the rules governing the designation of bodies responsible for conducting conformity assessments.

Although the decree does not primarily apply to undertakings that develop or deploy AI systems, it has an indirect impact on all market participants that intend to place high-risk AI solutions into operation in Hungary.

Read more

Phasing Out of the AVDH Authentication Service – Key Considerations for Businesses

2025-11-18
Phasing Out of the AVDH Authentication Service – Key Considerations for Businesses

As of 1 November 2025, the service of authentication of documents based on electronic identification (commonly known as AVDH) was discontinued.
This service was originally introduced by Act CCXXII of 2015 on Electronic Administration (hereinafter: the “Electronic Administration Act”), which also made electronic administration mandatory for economic operators. However, Act CIII of 2023 (hereinafter: the “Dáptv.”) initiated the gradual phasing-out of the AVDH service. Pursuant to Section 119 (2) of the Dáptv., private individuals have been prohibited from using the service since 1 January 2025, while economic operators were permitted to use it only until 31 October 2025.

Read more

The Use of Artificial Intelligence in Legal Matters – Opportunities and Risks

2025-11-18
The Use of Artificial Intelligence in Legal Matters – Opportunities and Risks

Artificial intelligence (AI), and in particular generative models, have undergone explosive development in recent years. An increasing number of individuals rely on these systems for various informational, analytical, or even legal questions. There is no doubt that AI tools can be extremely useful: they provide rapid responses, are capable of processing vast amounts of data, and often offer easily comprehensible explanations.

At the same time, it is important to recognise that, due to the current functioning and limitations of AI systems, significant risks also arise — including within the legal domain. For this reason, it is advisable to use such tools consciously and with caution, and not to accept their output at face value, particularly where legal questions are concerned.

Read more