{"id":128172,"date":"2025-04-29T05:15:57","date_gmt":"2025-04-29T05:15:57","guid":{"rendered":"http:\/\/cryptospotters.net\/?p=128172"},"modified":"2025-04-29T05:15:57","modified_gmt":"2025-04-29T05:15:57","slug":"zurich-unis-secret-ai-experiment-manipulated-redditors-opinions","status":"publish","type":"post","link":"http:\/\/cryptospotters.net\/?p=128172","title":{"rendered":"Zurich Uni\u2019s secret AI experiment manipulated Redditors\u2019 opinions"},"content":{"rendered":"<p>Source: Cointelegraph.com NewsResearchers at the University of Zurich have been accused of undertaking an unauthorized four-month-long experiment on a Reddit board using artificial intelligence-generated comments to test whether AI could alter people\u2019s opinions.\u00a0<br \/>\nAs part of the experiment, AI-powered accounts faked a variety of personas, including a rape victim, a person opposed to specific social movements and a trauma counselor specializing in abuse, moderators of the r\/changemyview subreddit said in an April 26 post.<br \/>\n\u201cOur sub is a decidedly human space that rejects undisclosed AI as a core value,\u201d the moderators said. \u201cPeople do not come here to discuss their views with AI or to be experimented upon.\u201d\u00a0<br \/>\nThe researchers used AI to generate responses but attempted to personalize the replies based on information from the original posters\u2019 prior Reddit history, such as political orientation, gender, age, and ethnicity, according to a draft of the paper.<br \/>\nModerators of the r\/changemyview subreddit say researchers at the University of Zurich conducted an unauthorized experiment within their community. Source: RedditOver the four-month experiment, Zurich University\u2019s fake AI accounts posted 1,783 comments and received 137 deltas \u2014 a mark showing when another Reddit user acknowledged the account had persuaded them to change their opinion on an issue.\u00a0<br \/>\nThe researchers said in a post to the subreddit that all comments were manually reviewed before posting to ensure they met the community guidelines and to \u201cminimize potential harm.\u201d<br \/>\nThe university team argued that the potential benefits of this research substantially outweigh its risks. Source: RedditThe university team argued that the \u201cpotential benefits of this research substantially outweigh its risks\u201d because they have found that the large language models used in the experiment can be highly persuasive, surpassing all \u201cpreviously known benchmarks of human persuasiveness.\u201d<br \/>\n\u201cOur controlled, low-risk study provided valuable insight into the real-world persuasive capabilities of LLMs \u2014 capabilities that are already easily accessible to anyone and that malicious actors could already exploit at scale for far more dangerous reasons,\u201d the researchers said.\u00a0<br \/>\nAt the same time, the researchers said the experiment shows that distinguishing humans from AI still faces significant challenges because nobody in the subreddit discovered the AI bots during the entire experiment.\u00a0<br \/>\nThe r\/changemyview board has 3.8 million members and ranks among the top 1% of subreddits by size.<br \/>\nReddit considers legal options\u00a0<br \/>\nThe Reddit board\u2019s moderators said they were unaware of the experiment and have asked the university for an apology and to block the publication of the research.\u00a0<br \/>\nReddit might also be considering legal action, as the platform\u2019s chief legal officer, Ben Lee, said in a follow-up April 28 post that the experiment broke the site\u2019s user agreement and rules, and all known accounts associated with the University of Zurich research effort were now banned.\u00a0<br \/>\n\u201cWe are in the process of reaching out to the University of Zurich and this particular research team with formal legal demands,\u201d Lee said.\u00a0<br \/>\n\u201cWe want to do everything we can to support the community and ensure that the researchers are held accountable for their misdeeds here,\u201d he added.\u00a0<br \/>\nRelated: White House receives over 10,000 comments on AI development plan<br \/>\nIn its response to the subreddit, the University of Zurich\u2019s Faculty of Arts and Sciences Ethics Commission said it had investigated the incident and promised to coordinate better with test subjects in the future. The lead investigator for the project was also issued a formal warning.<br \/>\n\u201cThis project yields important insights, and the risks are minimal. This means that suppressing publication is not proportionate to the importance of the insights the study yields,\u201d the ethics commission said.\u00a0<br \/>\nThe University of Zurich did not immediately respond to a request for comment. Cointelegraph did not immediately receive a response to questions sent to an email address associated with the experiment.<br \/>\nMagazine: Bitcoin $100K hopes on ice, SBF\u2019s mysterious prison move: Hodler\u2019s Digest<a href=\"https:\/\/cointelegraph.com\/news\/zurich-uni-secret-ai-experiment-manipulated-redditor-opinions?utm_source=rss_feed&amp;utm_medium=rss&amp;utm_campaign=rss_partner_inbound\" target=\"_blank\" class=\"feedzy-rss-link-icon\" rel=\"noopener\">Read More<\/a><\/p>","protected":false},"excerpt":{"rendered":"<p>Source: Cointelegraph.com NewsResearchers at the University of Zurich have been accused of undertaking an unauthorized four-month-long experiment on a Reddit board using artificial intelligence-generated comments to test whether AI could&hellip; <\/p>\n","protected":false},"author":0,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[5],"tags":[],"_links":{"self":[{"href":"http:\/\/cryptospotters.net\/index.php?rest_route=\/wp\/v2\/posts\/128172"}],"collection":[{"href":"http:\/\/cryptospotters.net\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/cryptospotters.net\/index.php?rest_route=\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"http:\/\/cryptospotters.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=128172"}],"version-history":[{"count":0,"href":"http:\/\/cryptospotters.net\/index.php?rest_route=\/wp\/v2\/posts\/128172\/revisions"}],"wp:attachment":[{"href":"http:\/\/cryptospotters.net\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=128172"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/cryptospotters.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=128172"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/cryptospotters.net\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=128172"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}