Following Google (Alphabet), the tech giant Meta has also decided to simply ban political, electoral, and social issue advertisements in the EU due to the “unenforceability” and excessive complexity of the new EU transparency rules (TTPA). Consequently, as of October 6, it is no longer possible to run such advertisements on Facebook.
Within the Central and Eastern European region, the Hungarian governing party, Fidesz has been particularly sensitive to these restrictions, as social media platforms here – especially Facebook – did not just serve a supplementary function but acted as the primary political battlefield. For instance, before the 2024 European Parliamentary elections, Fidesz and pro-Fidesz organizations spent a total of EUR 3.3 million on political ads between April 7 and June 1. For comparison, the opposition – including all other parties combined – spent a total of EUR 627,147 for the same purpose.
It is important to emphasize that when we speak of “Fidesz advertisements,” we do not only mean the governing party but also the “fake civil” organizations and media outlets connected to them through a thousand threads. Thus, “gray zone” advertisers like Megafon were already complicating transparency even before the bans – while Meta’s Ad Library showed the spending, it was only the illusion of transparency.
Also, before the bans, Political Capital conducted a survey covering a nearly one-year period: they examined how much various political actors spent on Meta and Google advertising between December 2024 and September 2025 (the last month before the restrictions). Out of the total EUR 10,585,970 spent, pro-government actors “blew” EUR 9,196,536, which accounts for more than 85% of the total Hungarian spending.
The study also shows that as the implementation of the bans approached, Fidesz pumped up its advertising even further. By the end of September 2025, the amount spent on ads approached figures previously seen during the 2024 elections – though at that time, Hungary held not only EP but also local municipal elections. The final month was pushed with spending of practically 200-300 million HUF per week before the shutdown of political advertisements on Facebook.
Gaming the Algorithm
From October, regular advertisers on Facebook had to adapt to the new guidelines. Naturally, the simplest way would have been for everyone to respect the bans and not “beard” with Meta’s algorithm. However, for more resourceful political actors, respond to the situation differently. Many continued to advertise because Meta’s algorithm – partly due to its own imperfections – initially struggled to filter out “smuggled” advertisements. If someone wanted to advertise in the EU and happened to select the political ad label during submission, the filter simply blocked it immediately.
However, if they clicked something else – for example, posting an in ‘entertainment’ category – it would pass through the filter, and there was a chance that if the system did not find the topic to be political, it would allow it through. Thus, those who still wished to advertise on Facebook chose a strategy of re-categorization during the self-declaration process.
For the governing parties, who have practically operated on unlimited amounts of public funds, and especially for organizations close to them, this was not a major blow – if an advertisement was eventually taken down during later monitoring, there was no other sanction at that moment other than the content being deleted. When well-established politicians and political influencers were not advertising, the goal shifted to posting as much as possible.
The governing party’s director, Gábor Kubatov actually issued this as an order: according to him, for the “digital conquest,” everyone’s activity had to be increased over one-and-a-half times. Through the conquest of the digital space, Facebook groups were formed to reinforce and echo the governing party’s political views, generally led and maintained by the influencers and politicians – thus, their roles shifted to some extent as well. As alternative communication channels, more and more politicians created their own chat channels/chatbots throughout the year, which simulate natural, informal conversations and provide continuous “status reports” like newsletters.
The Flooding Strategy: From Paid Clicks to Automated Narratives
The alternative that provided the real “breakthrough” for political actors flouting the bans was an even higher level of practice in the “flooding technique.” Starting from November, they tried to gain “organic reach” through fake users with AI-generated pictures; these fake profiles spread narratives in comments and groups. Lakmusz identified these profiles forming a network: these accounts, created and operated from the same mold, flooded their own pages with Fidesz propaganda, shared each other’s content, and commented under them to generate as much traffic as possible for their posts.
Growing AI-based “entertainment websites” operate on a very similar principle. These often reached the same numbers as paid advertisements, usually with low follower counts but massive view counts. The reason these are launched on Facebook as regular pages rather than personal profiles is that they are subject to looser regulations and are administratively easier to operate in parallel. This type of “advertising” is cheap, fast, and untraceable, as it operates through funding methods even worse than before; it is impossible to track who runs which page, how many only one person manages simultaneously, or if there is central control.
This flooding technique – which is much more effective than microtargeting – fundamentally defined political communication: while people encountered plenty of ads before, the launch of these pages multiplied the content. In the case of videos, users naturally encounter a wealth of deepfake content in connection with AI use – it is worrying that manipulated content shapes our opinions and deepens our prejudices even when we are aware of the fabrication, especially if the visual stimulus confronts us repeatedly. In the long run, this phenomenon not only strengthens social polarization but also spoils overall public trust, as continuous uncertainty leads us to question even authentic sources.
You might ask: if AI videos and fake profiles cause the same destruction as banned advertisements, why does Meta stand idly? On paper, the solution already exists: the EU regulation that entered into force in August 2024 but is becoming mandatory only this summer (AI Act) mandates the clear labeling of artificial content to avoid disinformation. In practice, however, a watermark rarely stops manipulation, and effective action is paralyzed by the EU’s convoluted and protracted bureaucracy.
Although Hungary is in the lead regarding rule-breaking, the problem is European-wide. Yet Meta does not tighten its grip, for two prosaic reasons: bureaucratic slowness and business interest. Under the Digital Services Act, irregularities must be proven individually by the European Commission, which is a slow process; thus, sanctions are absent, and the platform feels no real compulsion.
Furthermore, one of Meta’s primary sources of income is advertising: any step that limits the presence of paying customers – even if they trick the algorithms – goes against its fundamental business interests. So, the current state remains: the rules have been created, but their enforcement is lost in the tech giant’s hunger for profit and the labyrinths of European legislation.
Consequences and the Look Ahead
The current political situation in social media is evolving dynamically right before our eyes, yet it most closely resembles the Wild West: the somewhat transparent advertising system has been replaced by a struggle for untraceable and manipulated organic reach. Although Meta’s ban on political ads was intended to serve transparency on paper, in practice, it has rather added fuel to the fire on the Hungarian political battlefield. This begs the question: how useful was the ban, really?
It also raises the issue of why there are no more serious sanctions against those bypassing the rules. Currently, Meta’s algorithm only goes as far as deleting content, which represents a negligible risk for actors operating with almost unlimited public funds. For Meta and the EU to reach an agreement, a fine-tuning of the TTPA would be necessary – one that does not make a platform’s operation impossible but sets an effective barrier against manipulative content.
Looking to the future, an even more aggressive campaign than ever before is (and can be) expected in Hungary. In the spirit of the “digital conquest,” paid advertisements are being replaced by artificially boosted activity and deepfake content, which erodes public trust even when we know it is fake. Although the ban affects the entire EU, the situation in Hungary is particularly severe due to extreme spending ratios and established influencer networks.
The question is no longer who spends more, but who can more cleverly “trick” the algorithm and flood the voters’ information space. If no real solution is born, the possibility of authentic information may be lost forever in the digital Wild West.