In response to high-profile errors, including instructing users to put glue on their pizza and claiming Barack Obama was Muslim, Google has announced that it will be scaling down the use of AI-generated answers in certain search results. The decision comes after users and search engine experts noted a significant decrease in the number of queries triggering AI answers compared to previous days. In a blog post by Liz Reid, head of search at Google, it was confirmed that the company will be making adjustments to the AI answers, referred to as "AI Overviews. " These changes include reducing the use of social media posts as source material, pausing some answers related to health topics, and implementing triggering restrictions for queries where AI Overviews were not proving to be helpful. Additionally, the company has made numerous technical modifications to enhance the system. This is not the first time Google has launched an AI product with much fanfare, only to scale it back due to issues. In February, the company blocked users from creating images of people using its AI image tool after facing accusations of bias. The tech industry is currently experiencing an AI revolution, with start-ups and major tech companies alike incorporating the technology into their products and seeking ways to monetize it. However, some of these tools are being released prematurely as companies compete to be seen as cutting-edge. Google, known for its significant contributions to AI technology, has been striving to retain its status as a key player in the industry.

At its recent I/O conference, the company made over 100 AI-related announcements, with the most notable being the introduction of AI-generated answers in search results for the majority of its users. While the AI technology reads and provides summarized answers based on websites in search results, publishers have criticized Google for potentially harming their traffic and content distribution. However, the rollout of AI answers faced issues, with journalists, search engine experts, and social media users noting problematic responses. Some were amusing, while others raised concerns. These errors extended to sensitive queries, including health-related topics. Although Google conducted extensive testing prior to the broader launch, numerous unforeseen situations were only revealed during the full-scale implementation, according to Reid.