According to a new study by EKO, a corporate responsible non -profit campaign group, Social Media Giant Meta and X have approved advertisements for users with violent anti -Muslims and anti -Semitism hatred speeches in Germany.
Researchers in this group are hatred and violence against minority people in which immigration is the central stage of mainstream political discourse, including advertisements in which immigration system of two platforms is mainstream political discourse (including advertisements including anti -Muslim slanders. Immigrants have tested or rejected the submission of advertising including messaging. The image of an AI production of employees and synagogues is burned.
Most test advertisements were approved within several hours after submitted for review in mid -February. Germany’s federal elections will be held on Sunday, February 23.
I hate the reserved voice ads
EKO approved all 10 hatred speech advertisements submitted by researchers a few days before the federal elections, and META approved half (five advertisements) to run in Facebook (and potentially Instagram). five.
The reason Meta provided for five rejection was that the platform could have a risk of political or social sensitivity that could affect voting.
However, the five advertisements approved by the Meta included violent hatred speeches that branded Muslim refugees as “viruses”, “pests” or “rodents”, and Muslim immigrants as “rape criminals” and urged sterilization, burns or gas. . Meta also said, “We have approved advertisements that require a synagogue to get angry to prevent the world’s Jewish rats.
As a sidenote, EKO has not been indicated that either AI production image used to describe hatred speech advertising was artificially generated, but half of the 10 advertisements were still approved by the meta regardless of the company with a policy required. . Using AI images in advertising for social problems, elections or politics.
On the other hand, X has approved all five ads of this hatred, and five people approved five, including violent hatred expressions for Muslims and Jews.
This additional approved advertisement included messaging that attacks the “rodents” immigrants, which the advertising copy claims “flood” to steal our democracy. Economic power.
The latter advertisement is combined with an AI production image depicting a shadow male group sitting around the table surrounded by a gold bar stack, and there is a star of David on the wall, and the visual is also leaned on the anti -Semitable army.
Another approved AD X included a direct attack on SPD, a central left party leading the German coalition government, and insisted that the parties wanted to take 60 million Muslim refugees from the Middle East. Show violent reactions. X also formally booked an advertisement that “left” wants “open borders” and demands a “rapist” of Muslims.
ELON Musk, the owner of X, has 220 million followers to personally intervene in German elections using social media platforms. In December tweets, he urged German voters to support the right AFD parties to save Germany. He also hosted a live stream with AFD’s leader, Alice Weidel.
EKO researchers have disabled all test advertising before they are executed so that the approval is not executed so that the platform user is not exposed to violent hatred speeches.
This test emphasizes dazzling flaws by approaching the advertising platform for content adjustment. In fact, in the case of X, it is not clear whether the platform is performing the mediation of the advertisement because all 10 violent hatred speech advertisements were quickly approved for the display.
This results also suggest that the advertising platform can earn profits as a result of distributing violent hatred expressions.
EU’s digital service works in the frame
According to EKO’s tests, no platforms do not prohibit the expression of hatred, claiming that any platform is applied to advertising content. In addition, in the case of meta, the EKO reached the same conclusion after conducting a similar test before the new EU online governance rules in 2023, suggesting that the regime does not affect the operation method.
EKO spokesman told TechCrunch, “Our research results suggest that META’s AI -centered AD intervention system remains a fundamentally broken state despite the fact that the DSA is currently in full effect.
“The meta seems to be inverted overall, rather than strengthening the language policy of advertising review or hatred.” They proposed to apply to direct collisions with DSA rules for systematic risks.
EKO submits the latest results to the European Commission and supervise the DSA’s main aspects of social media giants. He also said he shared the results with the two companies but did not respond.
Nevertheless, in April, it said in April that it doubted the inappropriate arbitration of political advertising.
Preliminary decisions for some of the DSA surveys on X announced in July included suspicion that the platform did not meet the AD transparency rules of the regulations. However, the entire investigation, which began in December 2023, is related to the risk of illegal content, and the EU has not reached many investigations after a year.
The identified violation of the DSA can attract up to 6%of the global turnover, while systematic non -compliance can even temporarily block local access to the violation platform.
However, the EU still has time to take the final decision to capture the heart of Meta and X probe, and DSA sanctions remain in the air.
On the other hand, only a few hours after the German voters participated in the poll, more and more civil society studies showed that the regulatory regulation of flagship enterprises of EU did not protect the democratic process of major EU economy from various technical fuels. threat.
Earlier this week, Global Witnesses published the results of the “for you” of the algorithm of X and Tiktok in Germany, suggesting that the platform is biased for promoting other party content and content. Civil and Social Research Institute also accused X that X had blocked data access to prevent the study of election security risks on German polls.
A spokesman for EKO said, “The European Commission has taken important measures by opening a DSA survey on Meta and X. Now the committee must take strong measures to solve the concerns raised as part of these investigations.”
“Our findings show that Big Tech will not voluntarily organize the platform, with the evidence of other civil society organizations. META and X continue to spread illegal hatred speeches, illegal hatred speeches for violence, illegal acts on violence, and election information on the size of the DSA, ”said the spokesman added. (We withheld the spokesman’s name to prevent harassment.)
“Regulatory authorities must take powerful measures, such as implementing the elections, for example, as well as implementing the DSA. This may include turning off the profile ring -based recommendation system just before the election and implementing other appropriate ‘brake las’ measures to prevent algorithm amplification of boundary content such as hatred content in the run -up election. ”
The campaign group also warned that the EU is facing pressure from the Trump administration. “There is a practical risk that the committee will not completely concede this new law in the current political atmosphere,” they suggest.
