Following a wave of criticism over some of the more bizarre AI-generated answers, Google has begun to scale back its use of AI Overviews technology in its search engine. A study by SEO firm BrightEdge found that the number of search results answered fell from 11% to 7% in June.
However, as reported by The Verge, Google spokesperson Ashley Thompson has refuted BrightEdge’s findings, calling the company’s methodology flawed. Thompson claims that the study doesn’t account for the difference between users who have enabled AI Overviews as part of Google Search Labs’ experimental features and those who haven’t. He says that users who opt in to AI Overviews in Search Labs will see them in more search results. BrightEdge, for its part, claims that it only tracked data from people who have enabled AI Overviews.
Recall that since the launch of the AI Overviews short answers feature in Google search, Google has faced a number of embarrassments related to ridiculous and even dangerous advice given by AI. For example, users were advised to spread glue on pizza and eat rocks. In response to criticism, Google assured that it is working to improve its technology, limiting the use of unverified user content in AI responses and adding tools to detect meaningless queries that artificial intelligence should not receive.
Despite the challenges, Google has no plans to abandon the AI Overviews feature. CEO Sundar Pichai said that users have responded very positively to AI responses. However, it appears that Google is not yet ready for mass adoption of this technology and is continuing to refine it. The reduction in the use of AI Overviews may indicate that Google is simply afraid of a negative reaction and loss of trust in its search engine. Especially since users may turn to competitors such as Microsoft, OpenAI, and Perplexity.
If you notice an error, select it with your mouse and press CTRL+ENTER.
Source: 3dnews.ru