Do not fall for ‘AI summaries’

A STUDY commissioned by the New York Times has found that the “AI summaries” offered by Google instead of search results are wrong around one time in ten. So if you were to rely on them for fact-checking the odds on you destroying your career would not be in your favour.

see caption

An artificial intelligence lying through its teeth in a courtroom: the judge, she is not amused - as slopped out by ChatGPT

Image probably not ©

So: don’t.

For details see the links below... and savour those headlines.

And to further drive the point home: enterprising humans have discovered that they can fool “AI summaries” into repeating falsehoods about deeply obscure subjects as “fact” by making a single post to a service such as reddit.com. This works even when the post concludes “Just kidding. I don't know. I made all that up”

That was a fake answer to the question “Why do the Japanese like their buns askew?”

More seriously: Scientists invented a fake disease. AI told people it was real. That's a 7 April headline in Nature, which is arguably as serious as it gets. Almira Osmanovic Thunström, a medical researcher at the University of Gothenburg, Sweden, and colleagues invented the skin condition “bixonimania” and uploaded two fake studies about it to a preprint server in early 2024. A preprint server posting is supposed to be advance notice of work that's undergoing peer review: it's not checked in any detail.

“Within weeks of uploading information about the condition, attributed to a fictional author,” writes Chris Stokel-Walker, “major artificial-intelligence systems began repeating the invented condition as if it were real.

“Even more troublingly... the fake papers were then cited in peer-reviewed literature. Osmanovic Thunström says this suggests that some researchers are relying on AI-generated references without reading the underlying papers.”

©

* How Accurate Are Google’s A.I. Overviews? nytimes.com 07/04/26
* Testing suggests Google’s AI Overviews tell millions of lies per hour arstechnica.com 07/04/26
* Analysis Finds That Google’s AI Overviews Are Providing Misinformation at a Scale Possibly Unprecedented in the History of Human Civilization futurism.com 08/04/26
* 'Artificial Intelligence' our coverage to date