Part of the AI Regulation News hub.
Hungary has launched a government website at mihivatal.gov.hu where the public can report AI-related abuses and complain about AI use that harms their interests
A complaints channel is a cheap thing for a government to build and an expensive thing for a company to ignore. Hungary now has one aimed squarely at AI labelling and AI customer service.
Bottom line: An announcement of a website, not an instrument. It binds nobody and changes no obligation. What it does is give members of the public a named government address for AI grievances.
Who this affects: Customer service directors and heads of contact centres in Hungary running AI assistants, Hungarian marketing and content teams publishing AI-generated media, and in-house counsel who need to know where a complaint about their AI will land.
Issue date: 7 August 2026. No deadline is stated and none applies; this is a launch announcement.
What changed: The site mihivatal.gov.hu is live. According to the announcement the public can report AI-related abuses there and lodge a complaint where the use of the technology harms their legitimate interests, and the site carries practical advice for users of AI solutions.
Analysis: The two duties the announcement chooses to spell out are the two most visible to a consumer: labelling AI-generated content, and telling customers in advance when an AI assistant rather than a real person answers the contact centre. Those are also the two easiest for a member of the public to notice and report.
Primary sources: kormany.hu announcement, 7 August 2026 (Hungarian)
- Instrument (EN)
- Government announcement: A government website has launched at mihivatal.gov.hu
- Authority
- Government of Hungary, kormany.hu; the announcement is sourced to the ministry for science and technology and reports minister Zoltan Tanacs in indirect speech
- Jurisdiction
- Hungary, referring to EU Regulation 2024/1689 (AI Act)
- Status
- Website launched and live
- Bindingness
- Not binding. It announces a channel and restates AI Act duties; it creates none.
- Issue date / next deadline
- Published 7 August 2026. No deadline.
- Channel
- mihivatal.gov.hu, for questions about AI, for reporting AI-related abuse, and for complaints
- Duties named
- Labelling of AI-generated content, and advance notice to customers where an AI assistant answers
- Primary source
- https://kormany.hu/hirek/kormanyzati-honlap-indult-mihivatalgovhu-cimen
What launched, and what you can do on it
The government announced on 7 August 2026 that a government website has launched at mihivatal.gov.hu to answer questions connected with artificial intelligence, and that complaints can also be made there.
Zoltan Tanacs, the minister for science and technology, said that on the mihivatal.gov.hu government platform the public can report AI-related abuses and lodge a complaint if the application of the technology harms their legitimate interests. The announcement adds that the site supplies practical and useful advice to users applying AI solutions.
That is the whole of what the announcement claims for the site. It does not describe a procedure, a timetable for handling a complaint, or a remedy.
The two duties Hungary chose to advertise
The announcement frames the launch against deepfakes, saying that as AI develops it is ever easier to produce lifelike but false deepfake content capable of causing serious harm to people and businesses alike.
It then names the duties. Under the European Union's artificial intelligence regulation, AI-generated content must be labelled, and customers must be informed in advance about AI-based customer service.
Businesses carry a similar duty, the announcement says, where a customer service enquiry is answered not by a real person but by an AI-based assistant: they must inform consumers of that in advance.
Why a complaints channel is the operative fact
Nothing here alters the legal position of a Hungarian business. The AI Act duties the announcement describes already exist, and a government web page cannot add to them.
What the page does add is a route. Until now a Hungarian consumer who suspected an unlabelled AI chatbot or an unmarked synthetic image had no obvious government address; now the government has published one and told the public to use it.
For a compliance function, the practical implication is about detection rather than doctrine. Consumer-facing labelling failures are the kind of defect a member of the public can see from the outside, and a public reporting channel raises the rate at which they surface.
What the announcement does not say
It does not identify mihivatal.gov.hu as an authority, describe it as a market surveillance body under the AI Act, or say which body reviews the complaints submitted through it.
It cites no article number, no Hungarian implementing statute, and no penalty. It gives no indication whether a submission is a formal complaint with procedural consequences or an information channel.
Anyone advising a client should treat those gaps as gaps. The site itself may answer them, but this announcement does not.
What we did not verify
We opened the kormany.hu announcement of 7 August 2026 in full, in Hungarian, and every fact and the pull quote above come from it.
We did not open mihivatal.gov.hu itself, so we cannot describe its complaint form, its guidance content, its scope, or the body that processes submissions. We also did not open Hungary's AI Act implementing legislation and did not check whether any authority has been formally designated in connection with the site.
We will not describe the site as an AI regulator, an authority, or an official Article 50 guidance instrument. The announcement calls it a government website and a government platform, and that is as far as the source goes.
Treat this as a detection event rather than a legal one. Hungary has told the public where to complain about unlabelled AI content and undisclosed AI customer service, and those are the two failures a customer can spot without any inside knowledge. If you run a Hungarian contact centre with an AI assistant, the disclosure that appears before the conversation starts is now the control most likely to be tested by someone outside your company.
Source File
https://kormany.hu/hirek/kormanyzati-honlap-indult-mihivatalgovhu-cimen
Open the kormany.hu item dated 2026.08.07 and confirm that it announces the launch of mihivatal.gov.hu, that it says the public can report abuses and lodge complaints there, and that it states the duty to inform consumers in advance where an AI-based assistant answers customer service enquiries.
Hasonlo kotelezettseg terheli a vallalkozasokat is, amennyiben az ugyfelszolgalaton nem valos szemely, hanem egy MI-alapu asszisztens valaszol a megkeresesekre, errol elore tajekoztatniuk kell a fogyasztokat. (A similar obligation falls on businesses where enquiries to customer service are answered not by a real person but by an AI-based assistant: they must inform consumers of this in advance.) ยท kormany.hu, 7 August 2026
FAQ
Is mihivatal.gov.hu a regulator?
The announcement does not say so. It calls the site a government website and a government platform for answering AI-related questions, reporting abuse and lodging complaints. It names no supervising body and cites no designating instrument, and we did not verify one.
Does this create new obligations for Hungarian businesses?
No. The announcement restates duties it attributes to the EU artificial intelligence regulation: labelling AI-generated content, and informing customers in advance about AI-based customer service. A website launch cannot add obligations.
What happens after someone files a complaint there?
The announcement does not say. It gives no procedure, no timetable and no remedy, and we did not open the site to check whether it explains any of this.
Which AI uses does the announcement single out?
Deepfakes and AI-based customer service. It opens by noting that lifelike but false deepfake content is increasingly easy to produce and can seriously harm both people and businesses, then turns to the labelling and advance-notice duties.
Related briefings
Sponsored Training
Practical AI training for regulated professionals, built around verification, documentation and a defensible process. See the courses.