00:03Wszyscy jest dziennikarzył połączenia.
00:05Zobaczcie, jaka jest dziennikarzyła.
00:08Zobaczcie, jaka jest dziennikarzyła.
00:11Zobaczcie, jaka jest dziennikarzyła w szkolejach w Szakcone-Analch.
00:12Zobaczcie, jaka jest dziennikarzyła w Iranu i China.
00:16Zobaczcie, jaka jest dziennikarzyła.
00:20Wszyscy się dziennikarzyła, nie ma zakończeniem tego,
00:26jak również, na te wszystkie są słowości,
00:27które nie są wstępujące w swoich innych ludzi.
00:29In reality, they're AI-generated,
00:31and researchers say they form part of a coordinated network
00:35that's produced almost 1,900 political YouTube videos.
00:40These German-language videos cover everything
00:42from the wars in Iran and Ukraine
00:44to China, the United States and German politics.
00:48Researchers at the German non-profit CEMAS
00:51identified 47 channels using these artificial presenters.
00:55Together, the channels attracted more than 66,000 subscribers,
01:00while the videos recorded 7.5 million views
01:03and almost 30,000 comments.
01:05Nearly 800 addressed German domestic politics,
01:09criticising Chancellor Friedrich Merz's government.
01:12And compared to previous content,
01:14these videos are getting more sophisticated.
01:16The thing is, with AI-generated content,
01:19it's getting better by the minute, actually.
01:22So if we were to talk to each other in three months,
01:24we would have an entirely new set of quality
01:27that we're looking at.
01:28CEMAS has not been able to identify who created the channels
01:32or determine whether the motivation was political,
01:35financial or a combination of both.
01:37But researchers warned that as the technology improves,
01:41networks like this could manufacture entire groups
01:44of apparently independent experts,
01:46making some political narratives appear more accepted
01:49than they really are.
Comments