Organizations that contributed to this investigation: Greece Fact Check, PagellaPolitica/Facta news, Demagog Czech Republic, Maldita, Ellinika Hoaxes, Correctiv, RTVE, 15min, Ostro, Pravda Poland, Demagog Poland, Fact Review Greece, Faktisk, The Journal
Meta is facing increasing pressure in Europe due to the spread of scams across its platforms, such as Facebook and Instagram, particularly because of the growing use of advanced techniques such as deepfakes, fake advertisements, and misleading links. Over the past year, European authorities and organizations have expressed concerns about the effectiveness of the company’s measures to tackle these phenomena.
Investments in cryptocurrencies, gambling with guaranteed returns, loans and credit cards, but also alleged medical products and devices that pose a serious risk to public health: these are the most common scams spreading on the platforms, according to evidence gathered by EDMO. But there are also cases of fake giveaways such as boxes of clothes for little to no money, and sometimes well-known banks that will allegedly give a prize if the user performs an action such as answering some simple questions.
In order to enhance the credibility of these posts, scammers often rely on fake celebrity endorsements, including deepfake videos or links to cloned news, trustable institutions and enterprises as well as supposed government websites. In the case of medical products, the posts often include fabricated testimonials featuring well-known personalities or supposed medical professionals, some of whom are depicted as urging viewers to abandon their prescribed treatments.
The reach of such posts is difficult to evaluate, but multiple fact-checking organisations of the EDMO network have noted their appearance on a daily basis. But what does Meta do? In the past months, the company has announced the implementation of artificial intelligence tools, in combination with moderators, to detect such scams. Advertising Standards are also present with the aim of protecting users from fraud, combining automatic and human review.
However, a recent report by Fact Review Greece has highlighted that important aspects of this review process remain unclear. For example, according to the Systemic Risk Assessment and Mitigation Report published by Meta in August 2024, the company claims to be investing in improving its automated models and that its team is working 24/7 to review reported content. Nonetheless, as highlighted by Fact Review, the report does not provide any evaluation of how accurate the automated assessment is or what percentage of user-flagged content is actually reviewed by humans.
Moreover, human resources do not seem to be fitting adequately to the extent of the company’s market. Indeed, according to the Facebook Transparency Report published by Meta under the Digital Services Act (DSA) and covering the period from April 1 to September 30, 2024, the total number of content moderators for the 24 official languages of the European Union is 5.548, for both Facebook and Instagram. Meanwhile, each month Facebook alone has on average 260.6 million active users in the European Union.
Note: Moderators for languages such as English, Spanish, and Portuguese also provide global coverage beyond the EU.
The content
As long as it has no restrictions, any page can run different content and advertise it to a massive scale. A significant example reported by Greece Fact Check is the one of the scam page “Stories”, created on February 16, 2025. According to the Facebook Ads Library (a searchable database for advertisements that are running across Meta products), the page has to this day run 2.100 advertisements mostly related to unapproved medical products. Although dozens were removed because they did not comply with Meta’s Advertising Terms, or were served by an account or page that was later disabled for not following Meta’s Advertising Terms, as of June 10, 2025, the page in question is still running dozens of advertisements.
Another relevant example is the page “The Social Nest”, created on April 17, 2025. By the time it was deactivated on May 27, the page had already 500 advertisement results: all of them promoted unapproved diabetes products, as well as a supposed product for joint problems.
This scam alone, relying on deepfake videos to deceive users, ran in multiple languages and targeted many EU countries: the advertisements of the page redirected to websites in German, Italian, French, Czech, Polish, Bulgarian, Hungarian, and Greek. Meanwhile, the sponsor of the ads was listed as “Fraschetta Gourmet”: not a real person, but apparently the name of a restaurant in Rome, Italy. Moreover, in the Greek version of the scam the posts redirected to a website which displayed the official logo of the Greek Ministry of Health, misleading users into believing the products had been officially endorsed. Advertisements as such exploit public trust and medical vulnerability, particularly targeting patients with chronic or serious conditions. Their wide distribution on Meta platforms, often under misleading sponsorship identities, poses questions about the enforcement of DSA provisions related to advertising transparency and risk mitigation.
Pages promoting scam products can often have a very credible attire. This is the wide-spread case of fraudulent e-commerce schemes that present themselves as long-established, family-run businesses. Examples of such scams were found by our fact-checkers in Poland and in Italy, with a series of fake shops allegedly promoting a “closing down sale”. As reported by Facta, sponsored posts from alleged jewellery and clothing stores – both Italian and international – appeared online announcing their permanent closure after years of business and the subsequent total liquidation of their merchandise. In the case reported by Pravda Poland, the liquidation sale of the alleged shop “Kowalski Moda Wasawa” entailed discounts up to 70%.
These types of scams often involve gift giveaways, as were the cases noted in Slovenia by Ostro of electronic items sold for very little money (2 to 3 euros). Similarly, an investigation by Demagog Poland revealed advertisements promoting the sale of cheap travel suitcases filled with any sort of goods, which were presented as lost luggages that travelers from Warsaw Chopin Airport or other airports did not claim. False promises of easy winnings and guaranteed returns were also promoted by scam ads involving gaming applications that imitate “Plinko”-style games, as highlighted in investigations by Greece Fact Check.
Other well-designed frauds, instead, involve financial advisors who claim to help people invest. In one case reported in Norway by Faktisk, a woman lost all her savings and took out a loan she didn’t really need while following the advice of this alleged advisor, who in reality was a fraudster.
The techniques
Although the products and services advertised by scams may differ across countries and potential users, the underlying strategy is always the same. In all cases reported by fact-checking organizations members of the EDMO network, scams involve a good deal of emotional manipulation that leads users to trust the posts promoting the advertisements. In cases reported by Demagog Poland and Facta, the posts advertising the products were often positively commented by other “scam-users” providing fake reviews and therefore “social proof” to the fraud, i.e. a psychological effect that causes us to decide what to do on the basis of what we think other people have done. These scams also tend to rely on a narrative of urgency, pushing users to fall for the trap with messages such as «do this before tomorrow or else you will not get the offer».
Manipulative tactics also involve the use of pornography, which lures users to click on a specific link, or deepfakes that represent celebrities and important figures (such as politicians and journalists) allegedly endorsing the products. The latter method is especially becoming prevalent in the past year, as highlighted by various fact-checking organisations when reporting to EDMO for this investigation (Ostro in Slovenia, Correctiv in Germany, Ellinika Hoaxes in Greece, The Journal in Ireland, Faktisk in Norway and Maldita in Spain).
One of these examples, detected by Greece Fact Check, is a deepfake video of Mark Zuckerberg claiming to donate half his fortune (€200 million) and promoting a scam through a gambling app. Other examples reported by RTVE, Spain’s largest state-owned public media company, involve the King and the Queen of Spain allegedly promoting financial products, or the Spanish footballer Lamine Yamal advertising an app of RTVE itself. Moreover, RTVE has highlighted the presence on Meta’s platforms of scams presenting deepfake videos that impersonate the channel’s news anchors.
Notably, the choice of figures is often tailored to the country being targeted, with local or nationally recognized individuals used to maximize trust and engagement. In Cyprus, as reported by Fact Check Cyprus, a scam promoting an investing platform relied on the image of politician and YouTuber Fidias Panayiotou, elected as an independent MEP in the 2024 European Parliament election. In Germany, Correctiv has pointed out a scam ad in which AfD politician Alice Weidel was allegedly promoting a profitable investment platform, and in Poland, during election campaign, Demagog has noted that scam ads of various Polish politicians flooded social media.
Such tactics (well spread on other social media as well) allow scammers to exploit cultural familiarity and public recognition, increasing the likelihood that users will believe the claims and engage with the advertised, and often dangerous, products. These elements contribute to the storytelling of the scam, which is also always visually appealing, interesting and supposedly trustworthy. As reported by the Lithuanian fact-checking organization 15min to EDMO for this investigation, many of these fraudsters present themselves as trusted brands. As detected by Facta, this occurred also in Italy, where Decathlon was allegedly giving away backpacks branded “The North Face” for 2 euros.
Aside from the emotional manipulation of the victims, another element that these stories have in common is that the pages and profiles advertising scams tend to have a fake localisation, i.e. they claim to be based in a country while they operate from another one. For example, in the case of the “lost luggages” of travelers from Warsaw Chopin Airport, in Poland, the money of the victims was actually going to Cyprus. In another case reported by Facta, involving fake auctions of Italian villas promoted on a Facebook public group, six of the group’s administrators supposedly resided in Nigeria.
The location of the scammers and of the scammed is a crucial aspect to consider: an investigation by Tjekdet has revealed that in the case of Denmark, Facebook users who click on the ads from a Danish IP address are directed to the scammers’ fake investment pages. Meanwhile, users who click on the same ads from a non-Danish IP address are redirected to harmless websites. This strategy is designed to help scammers cheat Meta’s control system and target a specific language. If the company’s moderators don’t use a Danish IP address when reviewing the reported content (as it often happens, being many Meta teams based in the US), «they have no chance» of spotting it, said IT security expert Peter Kruse to Tjekdet. The same strategy has been proved to be used in Norway, as reported by Faktisk, in Greece, as noted by Greece Fact Check, and most likely in other countries.
Similarly, Correctiv has found out that another strategy ensures that while users are redirected to the scam page when clicking on the ad link, automated programs are shown a different page, such as a shopping website. This, once again, makes sure that ads circumvent Meta’s automated review.
The reach of these scams is hard to evaluate, but in general, fact-checkers from the EDMO network have highlighted that individual posts never generate much reach. The key factor to the success of the scam is the massive presence of these posts, advertised simultaneously by different pages and profiles that coordinate themselves. On top of this, as outlined by Demagog Czech Republic, the algorithm powering one’s individual feed on social media plays an important role: if someone “Likes” this type of content once, they will be more likely to get exposure to similar content.
Meta’s policy vs practice
According to a report by Fact Review Greece, «Meta’s more than $100B annual revenue depends almost exclusively on running ads on its platforms Facebook, Instagram, Messenger, and WhatsApp».
Based on the Digital Services Act (Article 26, “Transparency of Advertising”; Article 39, “Additional Transparency for VLOPs/VLOSEs”), as a Very Large Online Platform (VLOP), Meta is required to clearly label all of these advertisements as “Sponsored Content,” to provide transparency on the identity of the ads’ sponsor, to maintain a public ad archive (Ads Library) for one year for each ad, to offer information on reach, targeting, and spending, and to provide researchers access to analyze the influence and intent of advertisements.
When reviewing advertising disclosures on Facebook, users are meant to be able to identify who paid for a given ad and on whose behalf it is shown. In practice, however, the information provided is often vague, and misleading.
For example, advertisements promoting unapproved health products or deceptive investment schemes may list innocuous or unrelated entities as the payer or beneficiary, such as “KUNG FU Traditional Arts”, “Fraschetta Gourmet”, or similarly obscure names. These identities provide no meaningful insight into the actual commercial or editorial interests behind the ad, making it difficult for users, regulators, or fact-checkers to hold anyone accountable.
In order to counter the expanding trend of scam ads, however, transparency would still not be enough. Meta’s ads moderation, in fact, also looks inadequate. Even in very blatant cases such as the ones mentioned in this investigation, scam ads are often approved by Meta, and user reports receive no effective response.
As shown by multiple experiments conducted by Greece Fact Check, the enforcement over a scam report depended on who sent it. Indeed, fact-checkers first reported the scams as if they were common users: in this first case, Meta’s reply was that the content did not violate the platform’s rules. In a second instance, fact-checkers submitted a report as an organisation part of the Trusted Flagger program (i.e. entities nominated under the DSA in order to counter the spread of illegal content online). In this latter case, Meta always removed the content. Given the inconsistency in the enforcement of actions to contain scams, Greece Fact Check has signaled the issue to the local Digital Services Coordinator (DSC), responsible for supervising and enforcing the DSA while coordinating the other competent national authorities.
Inconsistency in the Meta’s response has also been highlighted by other fact-checking organisations part of the EDMO network. Some of them (Demagog Czech Republic, RVTE in Spain, 15min in Lithuania, The Journal in Ireland) have noticed a lack of response on Meta’s behalf, and others have pointed out that it is unclear whether fact-checkers’ reports on scams are actually used to combat the problem or not. Even if in some other cases (Elllinika Hoaxes in Greece, Demagog Poland) fact-checkers have reported that the enforcement to take down scam content is rather quick when reported directly, the overall trend of scams on Meta platforms does not seem to decrease significantly.
Under the Digital Services Act, Meta is required to provide a transparent reporting process and user feedback within the EU, but there remains a need for better public education and more effective mechanisms for preventing, and removing harmful content.
The widespread circulation of scams on Facebook and Instagram highlights the limitations of Meta’s current content moderation and advertising review systems, especially in the face of increasingly sophisticated fraud tactics like deepfakes, cloned websites, and cross-border language targeting. The challenge is not just technological, but it is structural and systemic. Despite regulatory frameworks like the Digital Services Act pushing for greater accountability and transparency, enforcement gaps remain, allowing malicious actors to exploit the platform’s advertising tools at scale. To effectively safeguard users, especially in Europe, Meta should not only improve its technical detection systems but probably also commit more resources to ensure swifter intervention, stronger identity verification of advertisers, and more meaningful collaboration with authorities and fact-checking organizations.
Thanos Sitistas, founder and director of Greece Fact Check, and IFCN Advisory Board member
Lucia Bertoldini, Journalist at Pagella Politica/Facta and EDMO
Photo: Daniel Foster/Flickr