Part I — Situation overview

On 4 September 2026 two pieces of news arrived about the Military National Security Service (KNBSZ), the military intelligence service under the direction of the Minister of Defence. The first: the invitation of the National Security Committee appeared on the parliament’s website, according to which the committee would hear Attila Szuvák — who has been temporarily leading the service as deputy director general — as nominee for director general at a closed session on 9 September. The second: investigative journalist Szabolcs Panyi reported that the service is searching for HVG’s sources, after the weekly, citing internal sources, wrote that the nomination had raised professional concerns among the staff. According to the report, the KNBSZ may be gathering data on 20 active and former members of staff using tools whose application requires neither judicial nor ministerial authorisation. These could therefore also have been ordered by the person temporarily leading the service on his own authority. The press office of the Ministry of Defence replied that it had no knowledge of soldiers or KNBSZ officers having given information to any media outlet. The Minister of Defence said in a video that he had not detected any leak, and that anyone who knows of one should file a criminal complaint. Neither of them denied the fact of the data gathering; in response to a follow-up question from 444.hu, the ministry stated only that “what has been said so far should be regarded as authoritative”.

The question is not without precedent. On 17 July 2026 MIAK wrote about the ministerial authorisation regime for secret surveillance for national security purposes, when a freedom-of-information request revealed that almost every application had been approved over 16 years. This case, however, lies on a different axis. The July question was who should authorise surveillance. The present one is what happens to the tools that require no external authorisation at all — and what if they are turned to tracing a newsroom’s sources. Hungarian law already recognises source protection at one point: in criminal proceedings a journalist may refuse to disclose a source, and only a court can compel disclosure, under strict conditions. In the national security procedure there is no such demarcation, so the clash between source protection and the national security interest is today decided not by a procedural test but by the service’s own discretion.

MIAK’s reading: this is not a personnel matter, nor a police investigation. The KNBSZ is a defence body, not a law-enforcement body, and the National Security Committee is a consultative, not an appointing, body — the nominee’s suitability and the tracing of sources are two separate questions. The problem is structural: if a service can search for the sources of reporting about its own head without external control, then it decides as an interested party in its own case. This is a lack of guarantees even if the concrete procedure should ultimately prove to be lawful.

Part II — Foundations in the literature

The topic can be understood from four sources. Article 10 of the European Convention on Human Rights (ECHR, the Council of Europe’s 1950 human rights convention, interpreted by the Strasbourg court) allows freedom of expression to be narrowed only by restrictions prescribed by law and necessary in a democratic society. Article 13 requires an effective remedy; subsequent notification lies at the intersection of the two. Section 174 of Act XC of 2017 on Criminal Procedure already contains a three-element judicial test for the disclosure of a source; the proposal would essentially extend this domestic standard. According to Alexis de Tocqueville (French political thinker, author of Democracy in America), where the citizen cannot bring the agents of the state before a court, the liberty of the press is not one guarantee among many but the only one. Edward S. Herman and Noam Chomsky (an American economist and a linguist and social critic, respectively) show in Manufacturing Consent that the news feeds on easily accessible official sources, while a non-official source is more costly from the outset. If the risk of contact with them also grows, these sources fall silent. The detailed treatment of the literature — by author, with quotations — is in section 6.4 Literature in detail.

Part III — MIAK’s concrete proposal

MIAK proposes three measurable measures. None of them restricts the service’s counter-intelligence work, and none of them constitutes interference in the ongoing case; all three fill in a missing demarcation rule. Since under Article 46(6) of the Fundamental Law the rules on the use of intelligence tools belong to a cardinal act, the amendment requires a two-thirds majority. This exists in the current National Assembly, so the decision is a matter of political will.

3.1 A judge should authorise the tracing of sources in advance — with a three-step test (bill by 31 December 2026)

MIAK proposes that the Government submit a bill amending Act CXXV of 1995 on the National Security Services (Nbtv.), to be adopted by the National Assembly. The core is a demarcation rule. If the purpose or foreseeable result of a covert intelligence-gathering tool is the identification of a journalist’s source, then prior judicial authorisation is required, irrespective of the type of tool. This also applies to the tools that today are not subject to external authorisation — at this point the proposal goes beyond the general authorisation reform of July. The judge — a judge of the Budapest-Capital Regional Court designated for this purpose — should examine three conditions, modelled on Section 174 of the Criminal Procedure Act: (1) there is a concrete and serious national security threat, for example classified information reaching a foreign power or a threat to life; (2) the necessary information cannot be obtained in any other way; (3) the interest in averting the threat clearly outweighs the interest in the source remaining secret (see 6.4.2). The law should also state that reporting on the person or appointment of the head of the service, or criticism of the service’s operation, does not in itself constitute a national security threat. In cases that cannot be postponed, judicial approval should have to be obtained retrospectively within 72 hours; failing that, the data must be destroyed. This is a concrete, procedural application of the A6 checks-and-balances programme point, which relies on the existence of an independent judiciary under I4.

3.2 Subsequent notification and judicial remedy for those concerned (together with the 3.1 law, in force from 1 March 2027)

MIAK proposes that after data gathering aimed at tracing sources has ended, the service should notify in writing, within 12 months at the latest, the journalist and newsroom concerned, as well as the persons about whom it gathered data in the course of searching for the source. In the present case the latter is the more significant group: according to reports, it concerns 20 active and former soldiers, who today cannot know whether data was gathered about them. Notification may be postponed only by judicial decision, with reasons, and in steps of no more than six months, for a total of no more than five years. The notified person should be able to turn to a court to have the lawfulness of the data gathering established and the unlawfully obtained data deleted. This is not the disclosure of the service’s secrets — the notification communicates the fact and the legal basis, not the operational details — but ensuring that a violation of rights has a remedy. Under Article 13 of the Convention the remedy is due even if the harm was caused by a person acting in an official capacity (see 6.4.1); without notification, however, the right to a remedy exists only on paper. The proposal complements the logic of the A5 whistleblower system: protecting the whistleblower means not only a ban on retaliation in the workplace, but also that the state should not be able to track him or her down in secret.

3.3 Mandatory register, annual statistics and clarification of the present case (immediately, and by 31 May 2027 respectively)

MIAK proposes that the service record in a separate register every use of a tool that concerns a journalist or the tracing of a journalist’s source, and report it to the National Security Committee within eight days. In its annual report the committee should publish — without revealing any specific case — how many such authorisations were requested per service, how many the court granted and how many it refused, and how many notifications were postponed. The first such summary should appear by 31 May 2027. As regards the present case, MIAK is asking not for a verdict but for clarification. The National Security Committee should examine, within its competence, whether data gathering aimed at tracing sources took place, who ordered it and on what legal basis, and publish its conclusion — without the classified details. This dispute is currently being conducted between assertion and silence; a committee finding could close it, whatever its outcome. The data in the register could also feed the A7 media-pluralism indicators and the annual media transparency report under KU2.

The common principle of the three proposals is that source protection should be entrusted not to the goodwill of the service but to a procedure. In criminal proceedings this standard already exists; nothing justifies weaker protection in a secret procedure without a remedy than in a public, judicial one. Tocqueville’s argument (see 6.4.3) warns of exactly this: where there is no accountability before a judge, the press remains the only check. And if the press’s sources can also be searched for in secret, then the last control is weakened too.

Part IV — Expected effects and risks

Dimension Expected effect Risk
National security Genuine counter-intelligence cases (foreign intelligence, leaks of classified information) can still be investigated with judicial authorisation; the service’s credibility grows Judicial authorisation may slow down urgent cases; without strict adherence to the exceptional, 72-hour retrospective approval, a loophole may arise
Press freedom The risk for staff who may come into question as sources becomes predictable; the flow of information in the public interest is not blocked If the court authorises as a formality, the guarantee may become a rubber stamp — as the July data series showed for ministerial authorisation
Judicial system The criminal-procedure and national security source-protection standards become uniform The caseload of the designated judges and the handling of classified documents require new capacity
Public-law structure The situation in which the service decides as an interested party in a matter concerning reporting about its own head comes to an end The definition of “purpose of tracing sources” may be disputed; if the service states a different purpose, the rule can be circumvented

The main question to weigh is how to distinguish legitimate counter-intelligence work from defence against inconvenient publicity. The proposal makes not the intention but the purpose and the foreseeable result the yardstick: if on any reasonable forecast the tool leads to the identification of a source, judicial authorisation is required. The system tips towards the risk side if judicial authorisation becomes a formality — which is why the refusal statistics under 3.3 are needed, showing year after year whether the judge actually filters. The other risk is the gaming of the stated purpose: this is counterbalanced by subsequent notification, because the person concerned can contest before the court what the purpose of the data gathering really was. An important boundary is that the proposal does not prohibit the investigation of the unlawful disclosure of classified information. If there is a suspicion of a criminal offence, there is a criminal-procedure route for it, where the judicial guarantees of source protection already apply today.

Part V — Measurability and summary

5.1 What is worth following? (suggested KPIs)

MIAK suggests following the performance indicators below (KPIs, Key Performance Indicators). These are proposals, not government decisions:

  • Whether the Government submits the source-protection amendment of the Nbtv. by 31 December 2026, and whether it also covers the tools that today are not subject to external authorisation (yes/no).
  • The court refusal rate of authorisation requests aimed at tracing sources in the first full year — a rate close to zero would indicate that the guarantee has become a formality.
  • The proportion of postponed notifications and the average length of postponement: postponement should be the exception, not the rule.
  • Whether the National Security Committee publishes by the end of 2026 its finding on the present case, not containing classified details (yes/no).

5.2 Summary

MIAK’s request is clear: data gathering aimed at tracing a journalist’s source — by whatever tool — should only be possible with prior judicial authorisation, on the basis of the three-step test known from criminal procedure, with subsequent notification and public annual statistics. And in the present case the National Security Committee should clarify publicly what happened. The request concerns neither the person of the nominee nor the government: MIAK would ask for the same guarantee under any government and any head of service.

Two MIAK foundational values are affected. Accountability, because covert tools by their nature fall outside public scrutiny; judicial authorisation and subsequent notification are precisely what allow accountability to take place without breaching secrecy. And transparency, because source protection is not a privilege of journalists but a condition of the public’s access to information: if the internal whistleblower can be identified, he or she stays silent, and the only information about the functioning of state institutions that reaches the citizen is what those institutions communicate themselves.

Part VI — Reasoning and further sources

6.1 Press framing by spectrum

The left-liberal band carried the story, and for a particular reason: HVG is both reporter and party concerned, since the case is about the paper’s own sources. HVG covered the case in two articles. In the first it presented the minister’s video and highlighted in the headline too that the head of the ministry did not respond regarding the tracing of sources; in the second it published Szabolcs Panyi’s conclusion, according to which the minister indirectly admitted that he had information on who could and who could not have been the sources. 444.hu carried the same conclusion and incorporated the ministry’s follow-up reaction, which gave no substantive answer. The same outlet also gave space to the criticism of former MP Ákos Hadházy, which recalled the minister’s earlier affairs. Telex was more restrained: it reported the nominee’s professional background and the official responses, and indicated that it had modified the title and content of the article after publication. The common emphasis of the band is press freedom; the legal regulation of intelligence tools remained in the background.

The public-affairs and business band carried another step by the defence ministry on the same day: 24.hu and Portfolio wrote about the appointment of a ministerial commissioner responsible for defence procurement, without the question of tracing sources. This is a signal in itself: the institutional restructuring of the defence sector appears as news, the question of oversight guarantees does not.

The conservative band brought up the topic through Magyar Nemzet, in a different frame: the paper described the minister’s “NER-talk” as cognitive dissonance and turned attention to the person of the journalist. Citing an audio recording made public in March, it raised the journalist’s alleged links to foreign intelligence services; according to the paper too, the then foreign minister concerned rejected these claims. The framing poses the question of whether someone who is himself accused of such links can credibly object to surveillance. By MIAK’s yardstick this question does not change the substance: the source-protection guarantee is tied not to the person of the journalist but to the source and to the public. It is precisely in such disputed cases that the cleanest procedure is needed. Mandiner did not cover the topic on that day.

6.2 Facts and data

Data Value Source
The alleged scope of the tracing of sources 20 active and former members of staff Szabolcs Panyi’s report, HVG and 444.hu, 4 September 2026
The nature of the tools used (according to the report) tools requiring neither judicial nor ministerial authorisation, on the head of service’s own authority HVG and 444.hu, 4 September 2026
The response of the Ministry of Defence it has no knowledge of soldiers or KNBSZ officers having given information; it did not deny the data gathering Telex and HVG, 4 September 2026
The committee hearing of the nominee for director general 9 September 2026, closed session HVG and Magyar Nemzet, 4 September 2026
The conditions for compelling disclosure of a source in criminal proceedings an intentional criminal offence punishable by at least 3 years’ imprisonment, the evidence cannot be substituted by other means, the interest in detection clearly outweighs source protection — by court decision Act XC of 2017, Section 174(2)
The constitutional basis of press freedom “Hungary shall recognise and protect the freedom of the press” Fundamental Law, Article IX(2)
The regulatory level of intelligence tools cardinal act (two-thirds majority) Fundamental Law, Article 46(6)
The governing party’s seats / the two-thirds threshold 141 / 199 seats (70.85%); threshold: 133 National Election Office (NVI), 19 April 2026
The Strasbourg court’s leading source-protection judgments Goodwin v. the United Kingdom (1996); Sanoma Uitgevers v. the Netherlands (2010); Telegraaf Media v. the Netherlands (2012) European Court of Human Rights
EU rule Article 4 of the European Media Freedom Act (EMFA, (EU) 2024/1083) on the protection of journalistic sources Official Journal of the European Union

Two rows of the table together give the essence of the proposal. In criminal proceedings a three-element judicial test is required to disclose a source; in the national security procedure, according to reports, tools can be deployed that require no external authorisation whatsoever. The difference between the two cannot be justified by national security cases being more serious. A serious case easily meets the judicial test — the test filters out precisely the cases that are not serious but inconvenient.

6.3 Policy dimensions

  • Defence (programme points) — the HV11 strategic communication and information protection programme point puts defence against hybrid threats, including foreign military intelligence, at the centre; the source-protection rule strengthens this, because it channels the service’s resources to the genuine counter-intelligence task and also protects the service’s public credibility. An important demarcation: the KNBSZ is a defence body, its operation is regulated not by the police act but by the national security act;
  • Justice (programme points) — the protection of judicial independence under I4 is the precondition for judicial authorisation being a genuine filter; and the I3 regulatory impact assessment requires, for the amendment of the Nbtv., an assessment of the impact on the service’s capacity and on the courts’ caseload;
  • Transparency and anti-corruption policy (programme points) — the strengthening of checks and balances under A6 provides the institutional framework of the proposal, A7 treats media pluralism as an institutional guarantee, and the A5 whistleblower system protects the same group — the internal whistleblower — from another direction;
  • Culture (programme points) — the annual report of the KU2 media-pluralism monitoring is the natural home for source-protection statistics, because it measures one of the conditions of editorial independence.

6.4 Literature in detail

6.4.1 European Convention on Human Rights

Article 10 of the Convention consists of two parts. The first paragraph states that freedom of expression includes the right to receive and impart information, without interference by public authority. The second paragraph lists the possible grounds for restriction — among them national security — but only if the restriction is

“prescribed by law and […] necessary in a democratic society, in the interests of national security”

The Strasbourg court derived source protection from this text in the Goodwin case (1996), and in the Sanoma case (2010) it also held that a judge or an independent body must decide before any interference affecting a source. In the Telegraaf Media case (2012) it was specifically an intelligence service that acted to trace a journalist’s source; the violation was grounded in the absence of prior independent control. Article 13 adds that everyone whose Convention rights have been violated is entitled to an effective remedy, “notwithstanding that the violation has been committed by persons acting in an official capacity”. Applied to the Hungarian case: the “prescribed by law” condition is vulnerable where no specific rule applies to the tracing of sources; and an effective remedy cannot be exercised without notification, since the person concerned does not even know that there is something to remedy.

📖 Source: Council of Europe: European Convention on Human Rights (Articles 10 and 13)

6.4.2 Act XC of 2017 on Criminal Procedure

Under Section 174 of the Criminal Procedure Act, a media content provider and its staff member may refuse to testify about who gave them information, unless the court has compelled disclosure of the source. The court may do so only if three conditions are met together: if knowing the identity of the source is indispensable for detecting an intentional criminal offence punishable by imprisonment of up to three years or more, if the evidence cannot be substituted by other means, and if the interest in detection is so outstanding that it clearly outweighs the interest in the source of information remaining secret. The act also makes the seizure of means of evidence held in a newsroom subject to a judicial decision. This standard is important because it shows that the Hungarian legislator has already accepted that source protection can be broken through only by judicial decision, on the basis of a strict test. In the national security procedure the same standard is missing — MIAK’s proposal 3.1 does not introduce a new principle, but extends an existing domestic principle to where the interference is secret, and where the guarantee is therefore all the more justified.

📖 Source: Act XC of 2017 on Criminal Procedure (Section 174)

6.4.3 Alexis de Tocqueville: Democracy in America

In his chapter on the liberty of the press in America, Tocqueville is surprisingly restrained: he confesses that he values the liberty of the press more for the evils it prevents than for the advantages it brings. At one point, however, he is categorical. There are countries — he writes — where certain agents of the state can violate the law with impunity, because those they oppress cannot bring them before a court:

“In this case the liberty of the press is not merely a guarantee, but it is the only guarantee, of their liberty and their security which the citizens possess.”

Tocqueville draws the conclusion that if the rulers of such a country were to narrow the independence of the press, the people could rightly demand in exchange to be able to bring the abuses of officials before the ordinary courts. Covert intelligence gathering is exactly such terrain. Someone who does not know they have been monitored cannot sue; so today only the public sphere — the press — exercises control over the covert tool. If the service is searching for the sources of precisely this control, both guarantees are harmed at once. From Tocqueville’s line of reasoning MIAK’s proposal takes over the judicial route, without narrowing the liberty of the press: prior authorisation and subsequent notification bring the covert tool, too, under judicial accountability.

📖 Source: Alexis de Tocqueville: Democracy in America

In their model of the “filters” of news production, Herman and Chomsky name sourcing as the third filter. The media are bound to the large institutional sources by economic necessity and reciprocity of interest: a newsroom needs a steady, reliable flow of raw material, which government and corporate bureaucracies regularly supply. The authors stress that material from official sources is cheap because it is regarded as credible from the outset, while information from non-official sources, which provokes criticism or threats, requires careful checking and costly research. The model was built for the American market environment, but the mechanism is general: the independent source is always more expensive. If on top of this a source can be identified by an intelligence service — without external authorisation — then the price of the non-official source also rises on the side of the internal whistleblower. The public is thus increasingly left with what the institutions communicate themselves. In the KNBSZ case this effect is independent of whether the concrete data gathering ultimately found anything: the mere possibility deters the next source.

📖 Source: Edward S. Herman–Noam Chomsky: Manufacturing Consent — The Political Economy of the Mass Media

6.5 International comparison

The most direct parallel is the Netherlands. The background to the Telegraaf Media judgment (2012) was that the Dutch civilian intelligence service used special powers against two journalists to find out from whom they had received classified documents. The Strasbourg court saw the violation not in the purpose of the service, but in the absence of prior independent control. The Netherlands has since introduced a specific rule in its intelligence services act: the use of powers aimed at tracing a journalist’s source requires the prior authorisation of the court in The Hague. This is exactly the structure that MIAK’s proposal 3.1 would fit into Hungarian regulation. In 2005 Belgium adopted a stand-alone source-protection act, which allows compelling disclosure of a source only by judicial decision and only to prevent crimes seriously threatening physical integrity. In 2010 France inserted source protection into its law on the freedom of the press, and allows it to be overridden only in the case of an overriding public interest. Article 4 of the European Media Freedom Act (EMFA), now applicable at EU level, makes surveillance and coercive measures aimed at tracing a source subject to case-by-case authorisation by a judicial or independent authority, justified by an overriding public interest, and requires effective judicial protection. Under the EU treaties national security remains a Member State competence. That is precisely why Hungarian rules must state that the same standard applies in the national security procedure too — if this does not happen, the guarantee will be weakest precisely on the most sensitive terrain.

Transparency and anti-corruption policy

  • A5 — Whistleblower system
  • A6 — Strengthening checks and balances
  • A7 — Media pluralism as an institutional guarantee

Justice

  • I3 — Regulatory impact assessment
  • I4 — Protecting judicial independence

Defence

  • HV11 — Strategic communication and information protection

Culture

  • KU2 — Media-pluralism monitoring

Proposed new programme point: Guarantee of journalistic source protection in national security procedures (prior judicial authorisation, subsequent notification, annual statistics) — for the Transparency and anti-corruption policy area, as a supplement to the A7 media-pluralism programme point, and as the source-protection counterpart of the general surveillance-authorisation programme point proposed in July 2026.

6.7 List of sources

Press sources (MIAK press monitor, 5 September 2026 — topic 5; MIAK topic monitor, 20 September 2026 — topic 1):

Knowledge-base references (literature and sources of law):

  • 📖 Council of Europe: European Convention on Human Rights (Articles 10 and 13)
  • 📖 Act XC of 2017 on Criminal Procedure (Section 174)
  • 📖 Alexis de Tocqueville: Democracy in America
  • 📖 Edward S. Herman–Noam Chomsky: Manufacturing Consent — The Political Economy of the Mass Media
  • Fundamental Law of Hungary (Article IX and Article 46) — as a source of law in the factual material

MIAK internal materials:

  • MIAK policy area: Defence (background material; programme points; programme point ID: HV11)
  • MIAK policy area: Justice (programme points; programme point ID: I3, I4)
  • MIAK policy area: Transparency and anti-corruption policy (programme points; programme point ID: A5, A6, A7)
  • MIAK policy area: Culture (background material; programme points; programme point ID: KU2)
  • MIAK legal foundations (background material) — the demarcation between defence and law-enforcement competences; the consultative role of the parliamentary committee
  • MIAK topic monitor, 20 September 2026 — topic 1, score: 100/100
  • MIAK press monitor, 5 September 2026 — topic 5, score: 88/100

Supplementary public data sources:

  • Act CXXV of 1995 on the National Security Services (National Legislation Database)
  • Act CIV of 2010 on the Freedom of the Press and the Fundamental Rules of Media Content — source-protection provision
  • European Court of Human Rights: Goodwin v. the United Kingdom (1996), Sanoma Uitgevers B.V. v. the Netherlands (2010), Telegraaf Media Nederland and Others v. the Netherlands (2012)
  • Regulation (EU) 2024/1083 of the European Parliament and of the Council (European Media Freedom Act), Article 4
  • Annual reports and session records of the National Security Committee (National Assembly)

Generation metadata

  • Input press monitor: MIAK topic monitor, 20 September 2026 (preceded by: MIAK press monitor, 5 September 2026)
  • Generation date: 22 September 2026, 06:14 CEST
  • Tokens used (total): ~95,000 — estimate, blog generation only; the token data of the source topic monitor is not available (see the tokens_breakdown field in the frontmatter)
  • Translation: Hungarian original at /blog/2026-09-22-ujsagiroi-forrasvedelem-elozetes-biroi-engedely-utolagos-ertesites/