- Google's AI search tool got simple mathematics wrong, a review by The Washington Post said.
- The experimental tool, Search Generative Experience, was introduced in May.
Google's AI search tool appears to struggle when it comes to simple mathematics.
The tool was put to the test by Washington Post reporter Geoffrey A. Fowler, who found that despite extensive public testing, it stumbled over a straightforward question.
Google previously said it was "supercharging" and "improving" its search experience with a generative AI-infused version called Search Generative Experience (SGE).
The AI question-answering tool is still an experiment. But as Fowler notes in a review published this week, it might have made it "dumber."
In the blog Google posted last May announcing the rollout, it said, "With new generative AI capabilities in Search, we're now taking more of the work out of searching, so you'll be able to understand a topic faster, uncover new viewpoints and insights, and get things done more easily."
Yet in Fowler's review, SGE fed back a response that was riddled with errors. The tech columnist searched for Meta CEO Mark Zuckerberg's net worth, and it replied, "$46.24 per hour, or $96,169 per year. This is equivalent to $8,014 per month, $1,849 per week, and $230.6 million per day."
The figures do not add up, which might be concerning given that Google is weighing up whether it should charge users for certain AI-powered features, The Financial Times reported.
Paying subscribers will rightly expect reliable answers, especially those focused on some of the world's most well-known business figures.
The company told the FT, "With our generative AI experiments in Search, we've already served billions of queries, and we're seeing positive Search query growth in all of our major markets. We're continuing to rapidly improve the product to serve new user needs."
Google already gives subscribers of its One AI Premium plan access to its AI model Gemini in features including Gmail, Docs, Slides, Sheets, and Meet.
But the company was forced to pull part of the Gemini AI model earlier this year after a series of blunders. Users complained the image-generating feature was creating historically inaccurate images of people of color.
Google didn't immediately respond to Business Insider's request for comment, made outside normal working hours.