What is the Rafał Brzoska vs. Meta dispute actually about?
Accountability for ads that use someone else’s likeness to run financial scams. For months, fake ads have circulated in Poland using the likeness of Rafał Brzoska, his wife Omenaa Mensah, and other public figures – supposedly endorsing investment platforms that don’t exist. Brzoska’s lawyers have identified over two hundred variants of this material. Despite an order from Poland’s data protection authority (UODO), which already told Meta to stop running these ads in Poland back in August 2024, and despite an April 2026 ruling by the Warsaw Court of Appeal rejecting Meta’s argument that it’s merely a “passive host” of ad content, the campaigns haven’t stopped.
As a result, Digitization Minister Krzysztof Gawkowski asked the European Commission to fine Meta €250 million for violating the EU’s Digital Services Act. Brzoska went further, proposing that any platform be fined 150% of the revenue it made from a specific scam campaign – a petition that over one hundred thousand people signed within a week.
What is ScamWatch, and why does it matter?
It’s a tool doing, for Meta, what Meta claimed was too technically hard. In late August 2026, Brzoska’s team launched ScamWatch – a platform built with the creators of Bielik, Poland’s own large language model. The system analyzes reported ads for deepfake artifacts, checks the infrastructure behind the linked domains, and clusters ads coming from the same criminal network. Once verified, the material is sent to the platform, and ScamWatch publicly counts how long its team takes to react.
That’s exactly the kind of transparency that should exist. Meta says it removed over 159 million fake ads globally last year, 92% of them before anyone reported them. ScamWatch doesn’t dispute that number – it just shows how long the reaction actually takes on a specific, documented report. Its creators are honest about what the tool is: not an enforcer, not a censor, not a court. It’s a public scoreboard designed to take away the platform’s “we can’t detect this” argument.
What does this dispute say about moderation in general?
That an advertising platform has a structural reason to delay its response. Scam ads, like toxic comments, generate something platforms monetize – attention. We’ve written about this before in the context of hate speech: a field experiment that automatically hid toxic content on Facebook, Twitter/X and YouTube found that when people saw less toxicity, they also spent less time on that platform and saw fewer of its ads – even if part of that attention simply shifted elsewhere online (we cover it in more depth here). The Brzoska dispute is the same mechanism at a larger scale and with higher stakes: a fake investment ad with a familiar face generates clicks in exactly the same way a hostile comment generates replies.
This isn’t about anyone at Meta wanting people to lose their savings or read insults. It’s that a business model built on ads and engagement has no built-in incentive to fix these things faster than it has to. It took the scale and determination of one of Poland’s wealthiest people to force a court precedent and a UODO investigation. That shows how high the bar has to be for a platform to react at all.
Why is hate in the comments the same story, minus the headlines?
Because it differs in the victim’s profile, not in the mechanism. Deepfake scams are expensive to produce and only target people with a recognizable face – billionaires, politicians, celebrities. That’s what gets them onto the front page, gives them a name, a law firm, and a shot at a legal precedent.
Hate in a comment section costs nothing to produce and doesn’t select its target by recognizability. It lands on the admin of a local animal shelter, a foundation volunteer, an independent journalist, the owner of a small business – anyone with the nerve to run a public conversation. We’ve written before about how much real damage this does to a brand’s or NGO’s reputation – organizations that don’t have the resources to wait for a court precedent, let alone hire a law firm to renegotiate the rules with Meta. This problem is fully democratized – and that’s exactly why no national newspaper covers it.
Does reporting content to the platform even work?
Depends how much patience and resources you have. If a billionaire with a legal team, state institutions behind him, and a public audit tool needs months to force a reaction to specific, documented reports, it’s hard to expect the “report” button under a single comment to work any faster for anyone else.
That doesn’t mean reporting is pointless – it is, especially for clear-cut violations. But manual moderation – whether that’s a report to the platform or a manual review by your own team – shares one weakness: it depends on someone else, somewhere else, finding the time. The reporting channel isn’t built for the scale and pace of a single page. It’s built for Meta’s scale.
How does ZenFeed help with this?
We built ZenFeed for exactly this reason – because waiting for someone else to react isn’t a strategy you can run a page on. The system analyzes every new comment for context, not a list of banned words, and removes or hides the ones that cross the threshold you set, in about 3 seconds – before anyone else gets to see them. Data is processed exclusively on servers in the EU.
Registered NGOs and independent newsrooms can use the full range of ZenFeed’s features for free through our non-profit program. Pricing details and limits are on the pricing page.
However Rafał Brzoska’s fight for Meta’s accountability over ads ends, keeping comment sections clean will stay in the hands of the people who run them. That’s why we believe calm public discourse is a right, not a privilege you have to win in court.