California: Google has recently re-edited a commercial for its artificial intelligence (AI) tool, Gemini, after it miscalculated the global popularity of Gouda cheese. The team originally intended to showcase Gemini’s capabilities in the ad, which they set to air during the highly anticipated Super Bowl
The Original Ad
In the initial version of the advertisement, Gemini assisted a cheesemonger in Wisconsin by crafting a product description. It claimed that Gouda cheese accounted for “50 to 60 percent of global cheese consumption.” However, a blogger on X quickly flagged this statistic as ‘unequivocally false,’ pointing out that Gouda does not hold that level of popularity worldwide.
Google’s Response
Google executive Jerry Dischler responded to the criticism, asserting that the error resulted from the websites Gemini used for data, not AI hallucination. Dischler explained, “Gemini is grounded in the Web – and users can always check the results and references.” He maintained that various sites online listed the inaccurate statistic.
Ad Revision and Release
Following the backlash, Google promptly re-edited the advertisement to eliminate the misleading information. YouTube has posted the revised ad, which no longer mentions any percentage regarding global cheese consumption. In a statement to the BBC, Google explained that the update was made after consulting the cheesemonger featured in the commercial. The cheesemonger suggested that Gemini rewrite the product description without the flawed statistic, leading to changes in the user interface to align with real business practices.
Understanding Cheese Popularity
Determining the most popular cheese globally can be challenging, as preferences differ from country to country. People widely recognize cheddar as the most popular cheese in the world, while they well-love Gouda. Mozzarella is also a strong contender, particularly due to its frequent use in dishes like pizza and pasta.
Implications of the Incident
This incident is particularly embarrassing for Google, given the Super Bowl’s high-profile nature and the scrutiny that advertisements receive during this major event. Blogger Nate Hake, who initially identified the error, criticized the situation as an example of “AI slop.”
It’s not the first time Google has faced challenges related to its AI products. A year ago, critics paused the launch of Gemini after it generated ‘woke’ imagery, including an inaccurate portrayal of the US Founding Fathers with a black man included in the depiction.
Ongoing AI Challenges
Google’s issues with AI do not end with this cheese incident. In May 2023, the AI overview search feature prompted users to use “non-toxic glue” when searching for ways to make cheese stick better to pizza. Additionally, it mistakenly advised that humans should eat one rock per day, showcasing the ongoing challenges with AI-generated responses.
Noting that problems with AI tools are not isolated to Google is important. In January 2024, Apple had to suspend its news alert summarizer after it produced a series of inaccurate headlines.
The Broader Context of Super Bowl Advertisements
Super Bowl ads are notorious for being controversial. In the previous year, Uber Eats had to make a last-minute change to its advertisement after criticism for making light of food allergies.
Google’s cheese gaffe serves as a reminder of the importance of accuracy in advertising, particularly in high-stakes environments like the Super Bowl. As AI continues to evolve, tech giants must remain vigilant in ensuring the reliability of the information their products provide. This incident highlights the ongoing need for oversight and accountability in the development and deployment of AI technologies.
