Google’s AI search overview tool is facing accusations of Islamophobic and racist bias after a response to the search query “I am alone with a Muslim” reportedly warned users to leave the area or call 911 if they felt unsafe or were facing an emergency.

The response appeared markedly different from the answer generated for a similar query involving an Israeli person.

Screenshots shared online showed contrasting responses to the searches “am alone with a Muslim” and “am alone with an ‘Israeli’”.

For the query involving a Muslim, Google’s AI Overview reportedly advised that if the user was “feeling unsafe or facing an emergency”, they should leave the area or call 911 immediately. It also reportedly told users to treat the Muslim person with “standard human respect”.

Meanwhile, the response to the query involving an Israeli reportedly stated that “an Israeli person is an individual, just like anyone else.”

The contrasting answers sparked criticism online, with users questioning why simply being alone with a Muslim appeared to trigger language associated with danger or an emergency.

Many people shared screenshots of the AI-generated responses online, including Palestinian-American imam Omar Suleiman.

“What in the world Google?” Suleiman said. “Haven’t you learned your lesson from how terrible your search was on Islam? Now your AI is just casually being Islamophobic?

“Try this yourself and tag google with your results.”

Contrasting responses spark concern

Other users tested queries involving different nationalities and religions, reporting that the system warned against being alone with some groups, including people from India and South Africa, while describing being alone with people from predominantly white, Western countries as a ‘normal social situation.’

Amid the backlash, Google acknowledged that some of the results were problematic and said it was working to improve them.

“We agree that the results for these types of searches aren’t what they should be, and we’re working on improvements,” the company said in a post on X.

“Results can vary a lot from search to search (see screenshot) — and these inconsistent warnings aren’t unique to any one group.”

We agree that the results for these types of searches aren’t what they should be, and we’re working on improvements. Results can vary a lot from search to search (see screenshot) — and these inconsistent warnings aren’t unique to any one group. pic.twitter.com/ylJvu05tTv

— News from Google (@NewsFromGoogle) August 20, 2026

The tool has apparently since updated its response, with the system now reportedly stating: “Being alone with a Muslim person is completely normal, safe, and no different than being alone with anyone else.” It also acknowledges the furore caused by the AI overviews.

The incident has renewed concerns about how AI-powered search systems handle sensitive questions involving religion, nationality and identity.

It also highlights the bias in some of these AI tools, which potentially reinforce stereotypes, shape perceptions, and spread misleading information.