Google has reissued an ad for its main artificial intelligence tool (AI), Gemini, after overestimating the world appetite for Gouda.
Advertising – which was supposed to present Gemini’s abilities – was created to be broadcast during the Super Bowl.
He showed that the tool helping a cheese in the Wisconsin write a product description by informing it that Gouda represents “50 to 60% of world cheese consumption”.
However, a blogger stressed on X that the statistic was “unequivocal” because the Dutch cheese was far from this popular.
Responding to him, the director of Google Jerry Dischler, insisted that it was not a “hallucination” – where AI systems invent false information – blame the websites that Gemini had reduced information to the place.
“Gemini is based on the web-and users can always check the results and references,” he wrote.
“In this case, several sites on the web include the stat of 50 to 60%.”
The announcement has now been reissued to delete the error.
Google published it on YouTube – which it has – and it no longer contains a reference to the percentage of the world consumes hard cheese.
In a statement, Google told the BBC that he had redone the announcement after talking to the Cheesemonger had presented and asking him what he would have done.
“Following his suggestion so that Gemini rewritten the product description without the statistics, we updated the user interface to reflect what the company would do,” added the declaration.
The precise characters on cheese are difficult to identify – but it is believed that cheddar is in fact the most popular variety in the world, although preferences vary from country to country. Another major competitor is mozzarella, given its use in pasta and pizza.
The incident is embarrassing for Google, given the extremely high profile of the Super Bowl and additional additional announcements to receive.
Blogger Nate Hake – who spotted the error – said it was an example of “slal ai”.
But this is not the first time that the technology giant has been found on the rear foot on its AI products.
A year ago, Gemini was “interrupted” following criticism of the “awakened” images she generated, as an image of the American founding fathers who included in an inaccurate manner a black man.
He even had previous problems with cheese.
In May of last year, its research function on AI previews was criticized after telling certain users to use “non -toxic glue” when they looked for `how to stick to cheese better with pizza ”.
The responses generated by the search engine ai-generation have also said that geologists recommend that humans eat a rock per day.
Problems with AI tools are not limited to Google – in January, Apple was forced to suspend its new alert summary after inventing a multitude of inaccurate titles.
The Super Bowl advertisements themselves are not foreign to controversy either.
Last year, Uber Eats brought a last minute change to his advertisement after criticizing that he shed light on food allergies.
