A made-up illness stemming from a 2017 hoax has made it into Google's AI search results. We'll explain to you what's behind it.
Just a few days ago, the Google AI Gemini made it into the headlines because it confronted a US student with a threatening message. The AI chatbot had the student and his sister saying “Please die. Please.” surprised.
But this isn't the only case of questionable results returned by Google's AI search. Because, as a Bluesky user writes, an illness he invented also made it into the search results.
Kyloren Syndrome: Invented disease in Google's AI search
The Bluesky user Neuroskeptic says he came up with the disease “Kyloren Syndrome” in 2017. He wrote on the platform that he invented the fictional illness “as part of a covert operation.”
He used the “Kyloren Syndrome” as the basis for a scientific article. His goal was to publish this article in a professional journal.
He wanted to prove “that you can publish any nonsense in greedy specialist magazines”. He apparently succeeded, because his fictional illness even made it into the results of Google's AI search.
Why doesn't AI recognize the fake?
Today he would have “thought twice” about the joke if I had known that I could contaminate AI databases with it, said Bluesky user Neuroskeptic in his thread. However, in 2017, this approach struck him as “a funny way” to draw attention to a scientific problem.
According to Google AI search, “Kyloren Syndrome” has become a supposedly real disease. This is caused by mutations in mitochondrial DNA and is passed on from mothers to their children
While Google's AI search presents the disease as fact, “normal” Google searches produce a different result. There is a PDF document there that clearly shows that it is a fake scientific paper.
However, in the “AI Overviews” the Google AI Gemini ignored this, even though the paper is cited as the source.
How does Google's AI search handle satire?
Wrong answers from AI chatbots can have dangerous consequences, especially when it comes to health issues. Google has also been criticized for this in the past.
In July 2024, research showed that “Google AI is providing false, potentially fatal information on various health questions” as answers, such as CBS News writes. The AI chatbot recommended eating “at least a small stone per day” so that the body receives all the important vitamins and minerals.
Since then, Google says it has restricted the inclusion of satirical and humorous websites in its health overviews. Viral search results have also been removed. But despite these measures, “Kyloren Syndrome” made it into the “AI Overview” as a supposedly real disease.
Also interesting:
Source: https://www.basicthinking.de/blog/2024/11/27/ki-suche-google-fake-krankheit/
