Last week, Google unveiled a significant change to its search engine, showcasing new artificial intelligence (AI) capabilities designed to answer user questions in an effort to catch up with competitors Microsoft and OpenAI. However, the new feature, known as AI Overview, has come under fire for generating numerous inaccuracies and errors, such as recommending glue as a pizza ingredient and suggesting the ingestion of rocks for nutrients. These mistakes have sparked online outrage and have damaged Google’s reputation as a reliable source of authoritative information.
The backlash underscores the heightened expectations placed on Google compared to other AI chatbots, which are also known for occasional falsehoods and odd behaviours. This pressure highlights the challenge Google faces in safely integrating AI into its search engine.
This is not the first time Google has encountered problems with its AI rollouts. In February 2023, the company’s chatbot Bard provided incorrect information about outer space during its announcement, causing Google’s market value to plummet by $100 billion. Earlier this year, Bard’s successor, Gemini, faced criticism for refusing to generate images of white people and inaccurately depicting historical figures.
Despite these setbacks, financial analysts argue that Google must continue to move swiftly to remain competitive, even if it means facing some initial issues. Thomas Monteiro, a Google analyst at Investing.com, noted that rapid development is crucial for the company, even if it results in some growing pains. “The user experience will just have to catch up,” Monteiro said.
Google spokesperson Lara Levin stated that the majority of AI Overview queries deliver “high-quality information,” with links for further exploration. She also noted that many problematic examples involved uncommon or doctored queries. Google plans to use these isolated cases to refine its system and improve the accuracy of AI-generated responses.
On Friday, Google announced it was taking “swift action” to enhance AI summaries of search results after users ridiculed several errors, including the claim that Barack Obama was the first Muslim US president. The company emphasized that it is using these incidents to implement broader improvements to its systems, some of which have already begun to roll out.
The erroneous suggestion to add non-toxic glue to pizza sauce, traced back to a child’s Reddit post, raised concerns about AI’s susceptibility to misinformation. While Google acknowledged these mistakes, it maintained that the vast majority of AI Overview responses are accurate and that safeguards are in place to prevent harmful content.
Google’s recent rollout of AI-generated answers to search queries in the United States represents one of the most significant changes to its search engine in 25 years. The AI summary now appears at the top of search results, with plans to extend this feature to other countries. This change comes as Google faces increasing competition from AI-powered search engines like Perplexity and rumours of OpenAI developing its search tool. Additionally, AI chat-based searches have appeared on platforms like Facebook, Instagram, and WhatsApp, offering users alternative ways to obtain information without relying on Google.
