AI Overviews under fire: Google's new search feature raises concerns about reliability

AI Overviews has been accused of giving inaccurate and sometimes humorous answers
An undated image of a person using a laptop. — Unsplash
An undated image of a person using a laptop. — Unsplash

Google's new artificial intelligence (AI) search feature, AI Overviews, has been criticised for providing random and inaccurate answers. The feature, which aims to provide a summary of search results, asked users to use "non-toxic glue" to improve the cheese on the pizza and recommended eating a "rock" a day.

The errors have gone viral on social media, raising concerns about the reliability of AI-powered search results. AI Review, which was launched in the US in mid-May, uses AI to provide a summary of search results.

However, it has been accused of giving inaccurate and sometimes humorous answers. For example, when asked if gasoline could be used to cook spaghetti faster, the AI suggested using gasoline to make a spicy spaghetti dish and provided a recipe.

Read more: Future of AI— Apple's WWDC set to unveil groundbreaking innovations

These errors, known as "hallucinations" in the AI world, have raised concerns about the reliability of AI-powered search results. Google has acknowledged the errors, saying they are "isolated instances" and not representative of most people's experiences.

The company has taken action where "policy violations" were identified and is using those instances to improve its systems. However, many are concerned that these errors may be more widespread than what Google is releasing.