Recently, Google introduced an AI feature called “AI Overview” that has been producing incorrect and potentially harmful summaries.
Recently, the search giant introduced the artificial intelligence function known as “AI Overview,” which has been providing users with summaries that are not only wrong but potentially harmful. Furthermore, it does not appear that Google has a solution to this issue.
Google To Restrict Certain AI Overview
At the time of writing, Google had restricted certain inquiries for its “AI Overview” function. Google made this decision due to widespread criticism of the system’s incorrect and potentially dangerous outputs. Reports began to emerge in the social media and news media communities about a user query that asked the search engine how to keep cheese on pizza.
The AI system purportedly responded with text stating that the user should use glue in order to maintain the cheese on pizza. In yet another instance of what appears to be a series of errors, the artificial intelligence system allegedly informed users that at least two canines owned hotels and referred to a dog statue that did not exist as evidence.
The primary concerns appear to be that the consumer-facing model that creates the “AI Overview” content produces inaccurate and accurate results with the same outward confidence. This is despite the fact that many of the supposedly inaccurate results appear to be comical or harmless.
Additionally, Meghann Farnsworth, a Google representative who spoke with the Verge via email, stated that the company’s task is to remove searches that produce erroneous results as soon as they appear. It seems that Google is tackling its artificial intelligence problem in a metaphorical manner.
To further complicate matters, it would appear that Google is placing the blame for the problems on the individuals who are responsible for producing the inquiries. Farnsworth notes that it’s unclear how users should refrain from asking “uncommon queries” at this moment.
Furthermore, it’s artificial intelligence system, typical of big language models, tends to generate different responses to the same questions upon repeated requests.
Despite the fact that it seems as though Google’s artificial intelligence system still requires some development in order to iron out the wrinkles, Elon Musk, the creator of xAI, a competing AI business, believes that these machines will surpass the capabilities of humans before the year 2025 comes to a close.
Musk shared his belief that xAI might catch up to OpenAI and DeepMind Google by the end of 2024 with attendees of the VivaTech 2024 event that took place in Paris.