One of Google’s AI Super Bowl ads is getting its facts all wrong

Google’s Super Bowl commercial featuring Gemini AI has drawn attention for making an inaccurate claim about global Gouda cheese consumption.

The questionable claim: Google’s advertisement showcasing Gemini AI’s capabilities across all 50 states included a misleading statistic about Gouda cheese consumption in its Wisconsin segment.

• The AI system claimed that Gouda makes up “50 to 60 percent of the world’s cheese consumption”
• This statistic appears to lack credible sourcing and contradicts expert knowledge about global cheese consumption patterns

Expert perspective: Cornell University’s agricultural economics expert provides important context about actual global cheese consumption patterns.

• Professor Andrew Novakovic clarifies that while Gouda may be commonly traded internationally, it is not the most widely consumed cheese variety
• He suggests that cheeses like Indian Paneer and various fresh cheeses from South America, Africa, and Asia likely have much higher consumption volumes

Source investigation: The questionable statistic’s origin raises concerns about AI systems’ data accuracy.

• The same statistic appears on Cheese.com, a website whose reliability has been debated on Reddit for over a decade
• Gemini does not provide a source for its claim, highlighting issues with AI transparency
• Professor Novakovic notes the absence of reliable global data on consumption patterns for specific cheese varieties

Commercial context: The advertisement’s presentation raises questions about AI reliability for business use.

• The ad shows a business owner using Gemini to generate website content
• While fine print states “This is a creative writing aid, and is not intended to be factual,” this disclaimer conflicts with the implied business use case
• Google recently integrated AI features into Workspace and increased subscription prices

Long-term implications: The inaccuracy in Google’s high-profile Super Bowl advertisement underscores ongoing concerns about AI systems’ reliability for factual content generation, particularly when used for business purposes where accuracy is crucial.