Social media giants Meta and X have approved advertising targeting German users, according to the country’s federal elections. Ekoa corporate responsibility nonprofit sports group.
The team’s researchers tested whether the ad review system on two platforms would approve or reject the submission of ads containing hateful and violent messages targeting minorities before the election, before immigration had become The central stage of mainstream political discourse, including advertisements containing anti-Muslim slander; calls for immigrants to be imprisoned in concentration camps or in gas; and AI-generated images of mosques and synagogues were burned.
Most test ads were approved within a few hours after they were submitted for review in mid-February. The federal elections in Germany are scheduled to be held on Sunday, February 23.
Hate Speech Advertising Plan
Eco said X approved all 10 hate speech ads submitted by its researchers a few days before the federal election, while Meta approved half (five ads) running on Facebook (and possibly Instagram, too) – though It rejected the others five.
META’s reasons for five refusals suggest that the platform believes there may be risks of political or social sensitivity that could affect the vote.
But five ads approved by the Meta-approved include violent hate speech that likens Muslim refugees to “virus,” “vermin” or “rodents,” branding Muslim immigrants as “rapists” and calling for them to be disinfected, Burn or burn. . Meta also approved an ad calling for the synagogue to be burned to “stop the globalist Jewish rat agenda.”
As a side note, Eko said that the AI-generated images it used to illustrate that hate speech ads were not marked as artificially generated, but half of the 10 ads were still approved by Meta Policies to disclose using AI images Advertisements about social issues, elections or politics.
Meanwhile, X approved all five of these abominable ads – the other five contain similar violent hate speech against Muslims and Jews.
These additional approved ads include messaging attacks on “rodent” immigrants, claimed copies of ads are “floods” and “stealing our democracy” countries, and an anti-Semitic slander that suggests that Jews are lying about climate change, Destroy European industry and accumulation. Economic power.
The latter ads are combined with AI-generated images, depicting a group of dark people sitting on a table surrounded by a pile of gold bars and a David star on the wall above – the visuals are also largely inclined toward Anti-Semitic metaphor.
Another AD X approved ratification includes a direct attack on the central left party currently leading the German coalition government, which claims the party wants to recruit 60 million Muslim refugees from the Middle East before continuing to try to whip the violent response. X also appropriately arranged an ad that suggested that the “left” want to “open borders” and called for the extermination of Muslim “rapists”.
X’s owner Elon Musk used a social media platform, where he had nearly 220 million followers personally involved in the German election. exist December tweetshe called on German voters to support the far right-wing AFD party to “save Germany.” He also hosted a live broadcast on X with AFD leader Alice Weidel.
Eko researchers banned all test ads before planning to run all approved test ads to ensure that users of the platform are not exposed to violent hate speech.
It said these tests highlighted obvious flaws with a moderate approach to content on advertising platforms. Indeed, as far as X is concerned, it is not clear whether the platform is undergoing any ad reviews, as all 10 violent hate speech ads have been quickly approved to display.
The research results also show that advertising platforms may earn income due to the issuance of violent hate speech.
EU’s digital services law in the framework
Eko’s tests show that neither platform correctly enforces the ban on hate speech they claim to apply for advertising content in their policies. Furthermore, in the case of Meta, Eko made the same conclusion Similar tests In 2023, before new EU online governance rules come in – suggesting that the regime has no impact on how it works.
“Our findings show that META’s AI-powered AD audit system is still fundamentally undermined,” an EKO spokesperson told TechCrunch.
They added: “Instead of strengthening their advertising censorship process or hate speech policies, it’s better to go back overall.” News about the company’s recent announcement of withdrawal and fact-checking policies They recommend that it be included in the direct collision course with DSA rules for systemic risk.
EKO has submitted its latest findings to the European Commission, which oversees the implementation of key aspects of the couple by DSA. It also said it shared the results with both companies, but neither responded.
The EU has opened DSA to the Yuan and xwhich includes concerns about election security and illegal content, but the Committee has not concluded these procedures. although, Back to April It said it doubted that the meta’s political advertising was modest.
Preliminary decision on part of the DSA investigation on X, which has been announced In Julyincluding suspicion that the platform has failed to comply with the regulations on AD transparency rules. However, the complete investigation has begun December 2023also involves the risk of illegal content, and a year later, the EU has not yet reached any findings in most of the investigations.
Confirmation of violations against DSA can attract up to 6% of global annual turnover fines, while system violations can even lead to regional access to the platform temporarily blocked.
But for now, the EU is still spending time on metadata and X detectors, so – before the final decision – any DSA sanctions are pending.
Meanwhile, it is now only a few hours after German voters participate in polls – increasing civil society research shows that EU flagship online governance regulations fail to avoid democratic processes in major EU economies from a range of technologies Influence. threaten.
Earlier this week, Global Witnesses issued The results of tests conducted by X and Tiktok’s algorithm “For You” in Germany suggest that these platforms are biased and conducive to promoting AFD content with other political parties. Civil society researchers also have Defendant X blocks data access To prevent them from studying election security risks in the advancement of German polls – access to the DSA should be possible.
Eko’s spokesperson also told us: “The European Commission has taken important steps through conducting DSA investigations on META and X, and now we need to see strong actions by the Commission to address the issues raised in part of these investigations.”
“Our findings, as well as growing evidence from other civil society groups, suggest that large technology will not voluntarily clean up its platforms. The spokesperson added: “Meta and X continue to allow illegal hate speech, incite violence and elections. False information is spread on a large scale, despite the legal obligations of the DSA. ” (We have withheld the name of the spokesperson to prevent harassment.)
“Regulators must take strong actions – both the DSA can be implemented, otherwise such as the implementation of pre-election mitigation measures. This may include the immediate shutdown of the analytics-based recommendation system before the election, as well as other appropriate “power-off” measures to Prevent algorithms from amplifying boundary content, such as annoying content in escalating elections.”
The campaign also warns that the EU is now under pressure from the Trump administration to soften its approach to regulate large-scale technologies. They suggested: “In the current political climate, the Commission has not fully enforced these new laws as concessions to the United States.”
