According to Google AI Overview, it gives users inaccurate results because of their requests to the AI.
“AI Overview,” a recently introduced artificial intelligence feature by search giant Google, has generated erroneous and potentially hazardous summaries in response to user queries. Google appears to have no concrete solution to this issue.
As of the composition of this article, Google has turned off specific queries for its “AI Overview” feature in response to extensive reports that the system was producing inaccurate and potentially detrimental results.
There was widespread dissemination of reports in the social and news media communities regarding a user inquiry to the search engine regarding the preservation of cheese on pizza. The AI system’s purported response was text advising the user to employ adhesive. In a subsequent series of apparent errors, the AI system allegedly advised users that hotels were owned by at least two canines, citing the absence of a dog statue as supporting evidence.
Although several purportedly erroneous outcomes may seem lighthearted or harmless, the primary issue appears to be that the consumer-facing model responsible for generating the “AI Overview” content confidently produces inaccurate and accurate results.
Furthermore, as stated in an email exchange with The Verge with Google representative Meghann Farnsworth, the organization has thus far been limited to promptly eliminating queries that elicit erroneous outcomes from the system. Regarding its artificial intelligence controversy, Google is engaging in a symbolic game of whack-a-mole.
Further complicating the situation, Google attributes responsibility for the problems to the individuals generating the queries.
As stated by Farnsworth:
“Many of the examples we’ve seen have been uncommon queries, and we’ve also seen examples that were doctored or that we couldn’t reproduce.”
The precise method by which users are expected to refrain from executing “uncommon queries” remains uncertain. Furthermore, like other sizable language models, Google’s AI system may produce inconsistent responses to identical queries when prompted repeatedly.
Cointelegraph attempted to obtain additional clarification from Google but has yet to receive a response.
Although Google’s AI system is currently undergoing refinement, Elon Musk, the founder of competing AI firm xAI, is confident that by the end of 2025, these machines will outperform human capabilities.
Musk recently told conference attendees at the VivaTech 2024 event in Paris that xAI could surpass OpenAI and DeepMind Google by the end of 2024, as reported by Cointelegraph.