Google's AI Overviews feature faces criticism for misleading responses
What's the story
Google's AI-generated search results, known as "AI Overviews," have recently been criticized for providing bizarre and misleading responses to user queries.
Critics have highlighted examples, such as suggesting the addition of glue to pizza sauce, and promoting the health benefits of tobacco for tweens.
These peculiar responses have led some critics to label the feature a "disaster" that "can no longer be trusted."
Expansion
AI Overviews expected to reach 1 billion users
The AI Overviews feature, which auto-generates summaries for complex user queries, was rolled out to all US users last week.
Even with ongoing issues that have damaged its credibility, it is projected to reach more than one billion users by the end of the year.
The feature has also been criticized for posing significant threats to traditional media outlets, by effectively demoting links to other websites.
Questionable advice
Bizarre suggestions and misinformation raise concerns
One notable example of the AI Overview's questionable advice was a response to a search query about cheese not sticking to pizza.
Google's AI suggested adding about 1/8cup of non-toxic glue to the sauce for added tackiness, a suggestion apparently based on an 11-year-old Reddit post.
Other examples include inaccurate information about which US presidents attended the University of Wisconsin-Madison and the potential health benefits of tobacco for tweens.
Company response
Google responds to criticisms
In response to the criticisms, a Google spokesperson stated that these examples were generally very uncommon queries and not representative of most people's experiences.
They added that their systems aim to prevent policy-violating content from appearing in AI Overviews, and will take action if such content does appear.
Despite these assurances, users have reported other instances where Google's AI provided peculiar answers.
Experimental phase
AI Overviews: Experimental and controversial
Google includes a disclaimer on all its AI Overview posts stating that "Generative AI is experimental."
The company has implemented safeguards on the system to prevent harmful content from appearing in results, and does not use AI Overviews on searches for potentially explicit or dangerous topics.
However, these mishaps are just the latest indication that Google's chatbot may not be ready for widespread use.