New amendments are being introduced to the legislation regulating the operation of social media platforms in Azerbaijan. Under the amendments adopted by the National Assembly, individuals under the age of 16 will be prohibited from creating social media accounts. Social media platforms will be required to verify users’ ages, establish a representative office or register in Azerbaijan, and respond to requests from government authorities within five working days. The amendments also provide for enforcement measures against platforms that fail to comply with these requirements, including fines, advertising bans, and the gradual reduction of internet traffic to their services.
At the same time, recent amendments to the Law on Information, Informatization, and Protection of Information have expanded the list of circumstances under which internet resources may be temporarily blocked without a court order. Under the new rules, content containing insults or defamation, information violating the right to privacy, the promotion of narcotic drugs, disinformation that poses a threat to public safety, and other similar types of content may serve as grounds for such blocking.
The authorities justify these amendments as measures to protect children, reduce risks in the digital environment, and strengthen information security. Critics, however, argue that the changes are part of a broader policy of increased control, following recent years’ actions such as the blocking of websites, the adoption of the Law on Media, the establishment of the Media Register, and the arrests of journalists.
What similarities and differences do these amendments have compared with international practice? How could the new mechanisms affect internet freedom?
On the issue, Arzu Geybulla, an Azerbaijani journalist and media expert known internationally for her reporting and commentary, spoke with Meydan TV and answered questions.
– In recent years, many countries have sought to restrict children’s use of social media. However, the new law adopted in Azerbaijan goes beyond simply imposing an age limit. It also requires platforms to establish a local representative office, respond to requests from government authorities, and, if they fail to comply, face financial penalties and even restrictions on their internet traffic. In your view, how does this model differ from approaches adopted internationally, and what is its defining legal characteristic?
– On paper, the amendments borrow from recognizable international templates — Australia’s under-16 minimum age, the EU’s Digital Services Act requirement for legal representatives, Germany’s NetzDG model of rapid removal of illegal content. The government has been careful to cite this “advanced international practice.” But regulation is never just the text of a law; it is the institutional environment in which the law operates. And that is where Azerbaijan’s approach diverges fundamentally.
In the EU, platform obligations are embedded in a system of safeguards: an independent regulator, judicial review, transparency requirements, user remedies. In Azerbaijan, enforcement is entrusted entirely to executive bodies — most recently the new National Cybersecurity Agency, created in June by presidential decree, which concentrates blocking, filtering, and personal data oversight in a single institution answerable to the minister and, ultimately, the president. There is no independent data protection authority, and in the nine years since the 2017 amendments introduced extrajudicial blocking, courts have not in a single documented case provided an effective remedy against a wrongful block.
The sanction toolkit is also telling. Advertising bans and the gradual throttling of a platform’s traffic appear in neither the Australian nor the European model — those instruments are lifted from Turkey’s 2020 social media law and Russia’s 2021 “landing” law, where they were used to force platforms into compliance with censorship demands. Fines, of course, exist in every model, but in Australia and the EU they are the end of the enforcement chain; here they are the prelude to degrading citizens’ access. And while some democracies do permit narrow administrative blocking — France, for example, for child abuse and terrorist content, under independent oversight and with appeal rights — the broad expansion of grounds for temporary blocking without a court decision has no equivalent in rights-respecting frameworks. So what we see is a familiar pattern: the protective vocabulary of child safety, and the enforcement architecture of information control.
“Local representative office gives the state a new tool of pressure”
– The new amendments require social media platforms to establish a local representative office in Azerbaijan, respond to requests from government authorities, and face sanctions if they fail to comply with these requirements. How are such mechanisms viewed in international practice?
– Mandatory local presence requirements are known among digital rights experts as “hostage-taking” provisions, and that is not rhetorical excess — it describes their mechanics. Once a platform has an office, staff, and assets inside a country’s jurisdiction, the state gains leverage it never had over a company headquartered abroad. Local employees can face legal liability, offices can be fined or raided, and this leverage is then used to extract compliance with content removal and user data demands. This is exactly how international human rights organizations assessed Turkey’s 2020 law — Freedom House called the legal-presence requirement an attempt to gain leverage and coerce platforms into censorship and surveillance — and the prediction was borne out: rights groups have since documented that after major platforms appointed local representatives, they became more willing to comply with removal orders than to challenge them. Russia’s 2021 “landing” law followed the same logic; platforms largely refused to comply, the state escalated sanctions, and after February 2022 most Western platforms were blocked or withdrew from the market altogether.
The obligation to answer state queries within five working days — formally tied to enforcing the registration and age-verification requirements, and backed by fines for “incomplete or distorted” information — deserves particular scrutiny. The law does not build in independent oversight of what state bodies may ask, and the compliance data these queries cover is, by design, identity data. In a country where people are routinely prosecuted for social media posts, a fast-track channel between state agencies and platforms is not an administrative convenience; it is a potential pipeline for identifying anonymous critics. The age verification system compounds this: tying account creation to bank cards and mobile numbers — which in Azerbaijan are registered to identity documents — builds a de facto identification layer over social media. Even with the law’s stated prohibition on storing this data, users have no independent authority to turn to if it is misused.
Arzu Geybulla, photo: Riccardo Urli
“This is not episodic censorship, this is institutionalized control”
– In recent years, Azerbaijan has introduced a media register, blocked news websites, arrested journalists, and is now moving to regulate social media. Taken together, what trend do these developments indicate for digital freedom of expression in the country?
– I see a consistent, cumulative project of closing every channel of independent information, one layer at a time. The 2017 amendments to the Law on Information created the legal machinery for blocking websites — and it was used within weeks against RFE/RL, Meydan TV, Azadliq, and others. When audiences and journalists migrated to social media and YouTube, the 2022 Media Law and its registry moved to define who counts as a journalist at all. The crackdown that began in November 2023 targeted the people — with around thirty journalists jailed, mostly on smuggling and tax charges that no serious observer considers credible. The Pegasus revelations showed the surveillance layer. And now the platform layer — the last remaining space where Azerbaijanis could access independent reporting and speak relatively freely — is being brought under regulatory control, at the same moment that a single new agency has been handed the technical keys to blocking, filtering, and data oversight.
Each of these steps was presented as something else: modernization, media professionalism, fighting smuggling, protecting children. Taken together, the direction is unmistakable — a shift from reactive, ad hoc censorship toward a permanent, institutionalized infrastructure of control. The trend is not deterioration; deterioration implies drift. This is construction.
Three possible scenarios for social media platforms
– What could be the long-term impact of this law on the operations of global platforms — such as TikTok, Meta, Google, and others — in Azerbaijan, as well as on users’ right to access information?
– There are essentially three scenarios, and none is good for users. If platforms comply, they open offices in Baku and become subject to the leverage I described — the predictable result, based on the Turkish experience, is faster removal of content the government dislikes and greater exposure of user data. If they refuse, they face fines, advertising bans, and throttling — meaning ordinary users experience degraded or blocked access to the main channels through which independent news reaches Azerbaijan. With their websites already blocked inside the country, exiled outlets like Meydan TV, Abzas Media, and RFE/RL depend almost entirely on these platforms for distribution — so throttling YouTube or Meta services would, in effect, finish blocking the independent press without a single new blocking decision. The third scenario — partial withdrawal of services — cannot be excluded either; Russia’s trajectory shows platforms do end up leaving, pushed out or walking away, when compliance costs exceed market value, and Azerbaijan is a small market.
For users, the long-term effects are chilling regardless of scenario: the erosion of anonymity through age verification infrastructure, self-censorship as people understand their accounts are traceable to bank cards and phone numbers, growing dependence on VPNs — which the state may then move against, as others in the region have — and a widening information gap between those with the skills to circumvent restrictions and those without. And an infrastructure built in the name of protecting children will not stay limited to children; verification and account-deletion mechanisms, once built, are available for any future purpose the state defines.
“The purpose of the law will become clear in its implementation”
– In your opinion, what indicators should be examined in the future to assess the effectiveness of these laws? For example, what criteria will be key in terms of platform operations, content removal, users’ rights, or judicial practice?
– First, enforcement selectivity: which platforms are sanctioned first, and whether enforcement actions cluster around politically sensitive moments — elections, protests, corruption revelations. If this law is genuinely about child safety, enforcement should not correlate with the political calendar.
Second, platform transparency reports: Meta, Google, and TikTok publish data on government removal and user data requests, though the detail of such reporting has been shrinking industry-wide. Watch whether requests from Azerbaijan spike after the law enters into force, and how compliance rates change — this is the most objective external measure we have, and it is especially telling here because Azerbaijan has historically appeared in these reports with very low request volumes, preferring direct blocking to asking platforms. Any surge will stand out against that near-zero baseline.
Third, technical measurement: network measurement data from OONI and similar projects will show whether throttling and blocking are actually applied, against whom, and by what method — especially important given the authorities’ practice of denying blocks, as we just saw with OC Media.
Fourth, court practice: whether a single platform sanction, blocking decision, or data demand is successfully challenged in an Azerbaijani court. Nine years of practice under the 2017 amendments have produced no documented case of an effective judicial remedy; if that continues, it confirms the law operates outside meaningful legal review.
Fifth, the fate of the age verification data: whether there are audits, breaches, or cases where information collected for age checks surfaces in prosecutions of users.
And finally, the government’s own stated goal: if this is child protection legislation, we should expect official reporting on measurable child safety outcomes. If the state measures compliance and sanctions but never measures child welfare, that tells us what the law was for.
