Content created by AI text generators is being mass-posted on dozens of websites as ostensible news articles in order to monetize the advertisements placed there. This is reported by the US company NewsGuard, which claims to have found 49 such portals in seven languages. The team apparently discovered this by searching for phrases such as “as an AI language model”, which are often found in the generated texts and were not removed from the clickbait articles. This also makes it clear that they are published without extensive human control. In one example, such a text module even ended up in the title – and is still online.
Masses of texts, inferior quality
The websites in Chinese, Czech, English, French, Portuguese, Thai and Tagalog have several things in common, NewsGuard summarizes. While it is usually not clear who is responsible, vast amounts of content are posted on a wide variety of topics, including politics, health, entertainment, finance and technology. Sometimes it’s hundreds a day. They are written “faintly” and full of repetitions. At the same time, the Internet pages are plastered with advertising, apparently money should be made with programmatic advertising. Answers to e-mails to the portals were sometimes “difficult to understand”, so they probably also came from ChatGPT & Co.
The pages found have generic names like “Biz Breaking News”, “News Live 79”, “Daily Business Post” or “Market News Reports”. Often the texts included clear credits as to authorship, such as the phrase, “I’m not able to produce 1500 words, but I can provide a synopsis”. Nevertheless, the inexperienced readers would often not realize that the texts are not written by people, writes NewsGuard. But the way it came about is most evident in one Text on “CountyLocalNews.com”, the title of which reads “Sorry, I can’t do this job because it violates ethical and moral principles”. The text is said to be about a police officer who is said to have died after being vaccinated.
(my)
