"I recently read that on 'Daily Time Update'," said hopefully no one ever. Websites like this are part of a never-ending flood of AI-generated content.
Media analysis company NewsGuard has uncovered a global network of 966 news websites that publish mostly AI-generated content in 16 different languages.
These "Unreliable AI-Generated News Sites" (UAINS) operate with little or no human oversight and publish articles that are predominantly or entirely written by bots. They have generic, more or less serious-sounding names like "iBusiness Day", "Ireland Top News", or "Daily Time Update" to appear like established news outlets.
To be classified as UAINS, according to NewsGuard, a significant portion of the content must be AI-produced without adequate human oversight. In addition, the use of AI must not be made transparent, so that the website appears to the average reader to be a news offering created by humans.
The AI-generated articles cover a wide variety of topics such as politics, technology, entertainment, and travel. They also repeatedly spread false claims, for example about politicians, events incorrectly presented as new, or alleged celebrity deaths. The sites are often funded through programmatic advertising, so well-known brands also unintentionally contribute to the monetization of these "botcontent farms".
AI has blooming fake news fantasies
In addition to the websites tracked in the UAINS tracker, NewsGuard previously found a Chinese government site that used AI texts to support the false claim that the US was operating a bioweapons lab in Kazakhstan that infected camels to endanger people in China. The company also identified 167 news sites with ties to Russia that pose as local media but spread AI-generated disinformation about the Ukraine war.
The researchers' findings seem particularly piquant because surveys have shown that people find it difficult to distinguish between human and AI-generated fake news. As an aggregator and the most important traffic provider, Google is also obligated to filter out low-quality content like this. However, the US company profits from content spam.
The company offers its database of UAINS domains domains as well as information on AI-supported disinformation to researchers, platforms, advertising customers and authorities.
