Google Search briefly declared OpenAI CEO Sam Altman dead after picking up a false edit on Wikipedia. The incident showed just how quickly misinformation can spread through AI-powered search.
It all started with a false edit on the open-source Wikipedia, which can theoretically be edited by anyone. The information soon spread across X and Reddit, creating an embarrassing situation for Google.
The search results stated that Altman was born on April 22, 1985, and died on August 12, 2026.
Shortly after the issue was brought to light, Google flagged the information as false and responded to a post on X: “Thanks for flagging. This is no longer appearing, and was not a manual change by Google
Notably, Google used the word “vandalism” in its response, stating: “When people vandalize public sources of information, it can affect the information that appears in Search.” By clarifying that the changes were not made manually by Google, the company effectively pointed to the actual source of the error.
As the modern search ecosystem increasingly shifts toward AI-powered solutions, greater attention is being paid to the accuracy of the content these systems collect and process.
Publishing nonsense on the internet is nothing new. However, AI analyzes both major sources, including Wikipedia, and lesser-known websites. This means that even if false information appears for only a few seconds, it can still be captured and spread, potentially putting major technology companies in an embarrassing position.
A study conducted in April found that Google’s AI Overviews have an accuracy rate of around 91%. While that figure may seem high, it also means the system could potentially produce tens of millions of errors every hour.
In June, a Munich court ruled that under German law, Google can be held liable for false claims made in its AI Overviews. Google’s AI had previously spread false information about several publishers, linking them to questionable business practices.
Another business was affected by a Google Search error in Japan. Earlier this month, AI Overviews falsely reported that a textile wholesaler had gone bankrupt, even though the company continued operating as usual.
That particular error originated from information in a paywalled Japanese newspaper article, which the system mistakenly associated with one of the company’s competitors that had actually filed for bankruptcy.
However, such hallucinations are not unique to Google. A similar incident involving Altman occurred in June, when DuckDuckGo’s AI incorrectly claimed that President Donald Trump had died from rabies.