It was a busy Memorial Day weekend for Google (GOOG, GOOGL), as the company scrambled to contain the fallout from a number of outlandish suggestions by its platform’s new AI Overview feature. research. In case you’re sunbathing on a beach or drinking hot dogs and beer instead of scrolling through Instagram (META) and X, let me bring you up to speed.
AI Overview is supposed to provide AI-based generative answers to search queries. Normally it does that. But last week users were also told they could use non-toxic glue to keep cheese from sliding off their pizza, that they could eat a stone a day and that Barack Obama was the first Muslim president .
Google responded by noting the responses and saying it used the errors to improve its systems. But these incidents, coupled with Google’s disastrous launch of the Gemini image generator, which allowed the app to generate historically inaccurate images, could seriously damage the search giant’s credibility.
“Google is supposed to be the premier source of information on the Internet,” said Chinmay Hegde, associate professor of computer science and engineering at NYU’s Tandon School of Engineering. “And if this product is watered down, it will slowly erode our trust in Google.”
Google’s AI failures
The issues with Google’s AI presentation aren’t the first time the company has encountered problems since beginning its generative AI program. The company’s Bard chatbot, which Google renamed Gemini in February, showed an error in one of its responses in a promotional video in February 2023, sending Google shares tumbling.
Then there was its Gemini image generator software, which generated photos of various groups of people in inaccurate settings, including German soldiers in 1943.
AI still has biases, and Google has tried to overcome this problem by including greater ethnic diversity when generating images of people. But the company overcorrected, and the software ended up rejecting some requests for images of people from specific backgrounds. Google responded by temporarily taking the software offline and apologizing for the episode.
The AI presentation problems, meanwhile, arose because Google said users were asking unusual questions. In the rock-eating example, a Google spokesperson said it “appears that a geology website is running articles on their site from other sources on that topic, and that includes a article originally appeared on The Onion. AI insights related to this source.
These are good explanations, but the fact that Google keeps releasing products with flaws that it then has to explain is getting tiring.
“At some point, you have to stick with the product you’re launching,” said Derek Leben, associate professor of business ethics at Carnegie Mellon University’s Tepper School of Business.
“You can’t just say…” We are going to integrate AI into all our well-established products, and it is also in constant beta mode, and we cannot be held responsible for any type of errors or issues that she trains. and even blamed, in terms of trust in the products themselves.
Google is the go-to website for finding information online. Every time I argue about something stupid with a friend, one of us inevitably yells, “Fine, Google it!” And chances are you’ve done the same thing. Maybe not because you wanted to prove that you know some obscure Simpsons facts better than your friend, but still. The fact is that Google has built a reputation for reliability, and its AI mistakes are slowly eating away at it.
A race to beat the competition
So why these slip-ups? Hegde says the company is simply moving too fast, releasing products before they’re ready in an effort to outmaneuver competitors like Microsoft (MSFT) and OpenAI.
“The pace of research is so fast that the gap between research and product seems to be narrowing considerably, which is the root cause of all these superficial problems,” he explained.
Google has been working to fend off appearances that it was lagging behind Microsoft and OpenAI since the two teamed up to release an AI-powered, generative version of its Bing search engine and chatbot. in February 2023. OpenAI even managed to exclude Google before its launch. /O developer conference earlier this month, announcing its powerful GPT-4o AI model a day before the show kicked off.
But if beating the competition means rolling out products that generate errors or harmful information, Google risks giving users the impression that its generative AI efforts are unreliable and, ultimately, not worth their time. to be used.
Email Daniel Howley at dhowley@yahoofinance.com. Follow him on Twitter at @DanielHowley.
Click here to know the latest technology news that will impact the stock market.
Read the latest financial and business news from Yahoo Finance
yahoo
