Google's AI-generated Super Bowl cheese ad has some holes
The AI-generated copy in the commercial claimed that Gouda accounts 'for 50 to 60 percent of the world's consumption.' An X user quickly pointed out that that's incorrect. ('Cheddar & mozzarella would like a word,' the user added.)
Though widely used since the release of OpenAI's ChatGPT in late 2022, generative AI still is prone to churning out inaccurate and oftentimes wacky results. Google's AI-generated search results came under fire when it provided misleading information, such as falsely claiming that former President Barack Obama is a Muslim. Other users pointed out last year the tool recommended adding glue on pizza to help the cheese stick.
Google is in the middle of a push to incorporate its Gemini AI technology across its suite of products in a race to keep up with its US AI competitors OpenAI and Meta, as well as new competitors in China.
On X, Google Cloud executive Jerry Dischler said that the gaffe wasn't a hallucination – which is when an AI model produces inaccurate or misleading results. Rather, he said 'multiple sites across the web include the 50-60% stat.'
In a statement, Google said it had updated the commercial. The statistic is erased from the commercial currently uploaded on Google-owned YouTube.
'After the question came up about the Gouda stat, we spoke with the owner of the Wisconsin Cheese Mart to ask him how he would handle. Following his suggestion to have Gemini rewrite the product description without the stat, we updated the UI to reflect what the business would do,' a Google spokesperson said Thursday.
When Googling 'gouda cheese consumption,' the first result that comes up is the 50-60% statistic from a site called Cheese.com. However, it's hard to find another website that uses that number, and the statistic has long been debated on the internet.
The Wisconsin ad was part of a series of 50 commercials from 50 different states' small businesses that show how they use Google Workspace with Gemini.
CNN's Clare Duffy contributed to this report.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
NBC News
9 minutes ago
- NBC News
Google embraces AI in classrooms with Gemini tools for students
As kids return to school, some teachers are embracing AI tools to assist them in the classroom. NBC News' Jesse Kirsch explains how a school is utilizing Google's Gemini AI tool to enhance learning for high school students. Aug. 20, 2025
Digital Trends
39 minutes ago
- Digital Trends
NASA and Google are testing an AI space doctor
With NASA eyeing long-duration crewed missions to the moon and beyond in the coming years, it has to be sure that if a medical situation arises, the astronauts are well equipped to deal with it. Currently, crews heading to the International Space Station (ISS) receive training for basic medical procedures and medicines, as well as for things like intravenous fluid administration, intubation, wound care, and basic emergency response. Recommended Videos But future missions that take humans hundreds of thousands — or even millions — of miles from Earth, potentially for years at a time, will add a new layer of complexity to health management. With that in mind, the U.S. space agency has partnered with Google on a project aimed at ensuring crew health and wellness on long-duration missions. The initiative includes an investigation into whether remote care capabilities can offer detailed diagnoses and treatment options when a health issue falls outside of the astronauts' knowledge base, and when real-time communication with Earth is limited. NASA and Google's work involves a proof-of-concept for an automated Clinical Decision Support System (CDSS) known as the 'Crew Medical Officer Digital Assistant' (CMO-DA). 'Designed to assist astronauts with medical help during extended space missions, this multi-modal interface leverages AI,' Google said in an online post. It said the CMO-DA tool could help astronauts 'autonomously diagnose and treat symptoms when crews are not in direct contact with Earth-based medical experts.' It added: 'Trained on spaceflight literature, the AI system uses cutting-edge natural language processing and machine learning techniques to safely provide real-time analyses of crew health and performance. The tool is designed to support a designated crew medical officer or flight surgeon in maintaining crew health and making medical decisions driven by data and predictive analytics.' Early results from initial trials have 'showed promise' for obtaining reliable diagnoses based on reported symptoms. Moving forward, NASA and Google are now working with medical doctors to improve the technology with a view to using it on future space missions. As part of NASA's Artemis program, astronauts could one stay for extended periods aboard a lunar satellite — similar to how they live and work aboard the ISS today — or even on the moon itself. More ambitious endeavors, such as to Mars, are also on the cards, but aren't expected to take place until the 2030s at the earliest.
CNET
an hour ago
- CNET
What Worries Americans About AI? Politics, Jobs and Friends
Americans have a lot of worries about artificial intelligence. Like job losses and energy use. Even more so: political chaos. All of that is a lot to blame on one new technology that was an afterthought to most people just a few years ago. Generative AI, in the few years since ChatGPT burst onto the scene, has become so ubiquitous in our lives that people have strong opinions about what it means and what it can do. A Reuters/Ipsos poll conducted Aug. 13-18 and released Tuesday dug into some of those specific concerns. It focused on the worries people had about the technology, and the general public has often had a negative perception. In this survey, 47% of respondents said they believe AI is bad for humanity, compared with 31% who disagreed with that statement. Compare those results with a Pew Research Center survey, released in April, that found 35% of the public believed AI would have a negative impact on the US, versus 17% who believed it would be positive. That sentiment flipped when Pew asked AI experts the same question. The experts were more optimistic: 56% said they expected a positive impact, and only 15% expected a negative one. Don't miss any of CNET's unbiased tech content and lab-based reviews. Add us as a preferred Google source on Chrome. The Reuters/Ipsos poll specifically highlights some of the immediate, tangible concerns many people have with the rapid expansion of generative AI technology, along with the less-specific fears about runaway robot intelligence. The numbers indicate more concern than comfort with those bigger-picture, long-term questions, like whether AI poses a risk to the future of humankind (58% agree, 20% disagree). But even larger portions of the American public are worried about more immediate issues. Foremost among those immediate issues is the potential that AI will disrupt political systems, with 77% of those polled saying they were concerned. AI tools, particularly image and video generators, have the potential to create distorting or manipulative content (known as deepfakes) that can mislead voters or undermine trust in political information, particularly on social media. Most Americans, at 71%, said they were concerned AI would cause too many people to lose jobs. The impact of AI on the workforce is expected to be significant, with some companies already talking about being "AI-first." AI developers and business leaders tout the technology's ability to make workers more efficient. But other polls have also shown how common fears of job loss are. The April Pew survey found 64% of Americans and 39% of AI experts thought there would be fewer jobs in the US in 20 years because of AI. Read more: AI Essentials: 29 Ways You Can Make Gen AI Work for You, According to Our Experts But the Reuters/Ipsos poll also noted two other worries that have become more mainstream: the effect of AI on personal relationships and energy consumption. Two-thirds of respondents in the poll said they were concerned about AI's use as a replacement for in-person relationships. Generative AI's human-like tone (which comes from the fact that it was trained on, and therefore replicates, stuff written by humans) has led many users to treat chatbots and characters as if they were, well, actual friends. This is widespread enough that OpenAI, when it rolled out the new GPT-5 model this month, had to bring back an older model that had a more conversational tone because users felt like they'd lost a friend. Even OpenAI CEO Sam Altman acknowledged that users treating AI as a kind of therapist or life coach made him "uneasy." The energy demands of AI are also significant and a concern for 61% of Americans surveyed. The demand comes from the massive amounts of computing power required to train and run large language models like OpenAI's ChatGPT and Google's Gemini. The data centers that house these computers are like giant AI factories, and they're taking up space, electricity and water in a growing number of places.
