
Google Admits AI Overviews Tool Needs Improvement, Scaling Back Features
Google's AI Overviews tool needs improvement due to generating odd and erroneous responses, leading to scaling back and ongoing improvements.

After the recent launch of its AI Overviews tool, Google has acknowledged that the artificial intelligence-powered feature requires significant improvement. Despite extensive testing before its release two weeks ago, the technology has been found to produce "odd and erroneous overviews", including suggestions such as using glue to make cheese stick to pizza or drinking urine to quickly pass kidney stones.
Rollback of AI Product
This setback represents the latest case of a tech company hastily introducing an AI product to establish itself as a leader in the competitive space. Due to the generation of unhelpful responses to search queries, Google has decided to scale back the AI Overviews feature while striving to make necessary enhancements. Liz Reid, Google's head of search, acknowledged the occurrence of "odd, inaccurate, or unhelpful AI Overviews," emphasizing the need for improvements in specific areas. She also highlighted nonsensical questions that prompted questionable content from AI Overviews, attributing this to the lack of useful advice on the internet.
Addressing Inaccurate Information
Reid revealed that the AI Overviews feature is susceptible to misinterpreting language on webpages and presenting inaccurate information in response to search queries. To rectify this, Google has promptly taken action by making improvements to its algorithms and implementing processes to eliminate non-compliant responses. In addition, the company has introduced triggering restrictions for queries where AI Overviews were deemed unhelpful and has limited the use of user-generated content that could potentially offer misleading advice.
Limitations and Updates
Google has announced its decision to refrain from displaying AI Overviews for hard news topics where freshness and factuality are crucial. Moreover, the company has made updates to mitigate the use of user-generated content in responses, thereby reducing the risk of offering misleading advice to users.
Share news