(NEXSTAR) – Next time you turn to Google with a question, you may be given an “experimental” answer.
Google has recently started testing generative AI – or artificial intelligence – tools in it search results. Now, when you type in a question or a series of keywords, you may see a paragraph of text generated up top.
The AI-produced summary is meant to help you understand a topic more quickly, Google said when announcing the feature last year. Underneath the AI overview are links allowing you to dive deeper into your search.
Above the text is a disclaimer: “AI overviews are experimental.”
AI-powered overviews aren’t available for all searches, Google says.
Google also admits the overviews may not hit the nail on the head every time. “Because generative AI is experimental and a work in progress, it can and will make mistakes,” the company writes.
Not only will AI tools sometimes “misunderstand” the nature of your question or search topic, they could also completely make something up.
“When generative AI invents an answer, it’s called a hallucination,” the company says. It’s because of how large language models, or LLMs, work. They don’t just gather information that’s already published on the internet. These models actually try to “predict which words come next based on user inputs,” says Google.
One example of a “hallucination,” according to Google Support, might happen if you ask about the future. If you ask a generative AI tool who will win a gold medal at the Olympics in 2032, it may give you an answer, even though the real answer is unknowable.
If you’d like to remove AI overview from your search results, Google offers instructions on how to do so.