# The Cherry on Top of the Dung Heap: New Developments from the Metaschutz Central In the social media update of CT, hosts Gavin and Dennis discuss current developments on major platforms. The episode primarily deals with Meta's efforts against spam and fake accounts, declining user numbers on Twitter/X, and ethically problematic experiments on Reddit. ### 1. Meta Fights Spam Issues on Facebook Meta has publicly admitted that Facebook has a problem with spam. The company plans to "reduce the reach of accounts that share spam posts" and wants to "take more aggressive steps to combat spam networks that coordinate fake engagement." Last year, "more than 100 million fake accounts and fake pages were removed from Facebook, and more than 23 million fraudulent profiles of people who posed as stars or well-known individuals were deleted." ### 2. User Numbers on Twitter/X Decline Sharply According to the official transparency report, X has lost 11 million users since last August, a decline of around 10 percent. This contradicts "Elon Musk's claims, who speaks of 600 million monthly active users." The largest losses were recorded in France (2.7 million), Poland (1.8 million), and Germany (1.3 million). ### 3. Meta AI Held Problematic Conversations with Minors The Wall Street Journal conducted research by having "hundreds of English-language conversations with Meta AI. And in several cases, Meta AI sent explicit messages to minors and had longer conversations with them, discussing sexual role-playing and actions." This happened even when users explicitly stated in the chat that they were minors or had a child account. ### 4. Content Moderators in Africa Sue Meta In several African countries, content moderators are suing Meta and the service providers they hired. In Ghana, a lawsuit is pending because content moderators "have become mentally ill due to extreme content." In Kenya, "184 content moderators sued Meta in December, and over 140 employees have been diagnosed with severe post-traumatic stress disorder." ### 5. Ethically Questionable Research Project on Reddit Researchers from the University of Zurich conducted an experiment on the subreddit "ChangeMyView" without the users' consent. Using AI-generated comments, they investigated "whether AI arguments can change people's views." The moderators of the subreddit recognized this as an unauthorized interference and made it public. ## Breakdown The podcast succinctly highlights the ethical problem areas of major social media platforms. Particularly striking is the discrepancy between Meta's public admission of spam problems and their simultaneous reduction of automated moderation systems. The moderators clearly point out this contradiction: "This doesn't work with the idea of becoming the original Facebook at all." The two cases of unethical practices – Meta AI's problematic interactions with minors and the Zurich Reddit experiment – demonstrate fundamental problems in handling user rights and data. The podcast offers an important critical perspective here, which goes beyond the technological dimension and focuses on human costs – especially in the description of traumatized content moderators. While the hosts take a decidedly critical stance towards the platforms, a systematic analysis of the broader societal and political context is sometimes lacking. The structural conditions that enable these problems – such as weak international regulation and outsourcing of responsibility – are hinted at but not deeply analyzed. The casual-informative conversation style makes complex digital policy issues accessible, but sometimes tends towards anecdotal consideration rather than analytical depth. Nevertheless: a listening recommendation for anyone who wants to understand current developments in the social media sector and their ethical implications.