"Taking Swift Action":Google On AI Overview Falsely Claiming Obama A Muslim

This unique feature has landed the company in trouble after it reportedly provided false or misleading information to some questions

'Taking Swift Action':Google On AI Overview Falsely Claiming Obama A Muslim

With its new artificial intelligence search tools, Google promised to "do the work for you" and make the search engine faster and smoother. However, it seems the global tech giant is facing problems within weeks after its launch.

Earlier this month, the US-based company unveiled an AI-generated search results overview tool. This one-of-its-kind tool is capable of summarising search results for the users, so they don't have to click through multiple links to look for quick answers.

This unique feature has landed the company in trouble after it reportedly provided false or misleading information to some questions, CNN reported.

To back this claim, several users on the social media platform, X, posted how the tech giant's AI summary stated that former US President Barack Obama is Muslim -- a common misconception about him.

This is not the only misleading information provided to users. Another one claimed that a Google AI summary stated that “none of Africa's 54 recognised countries start with the letter ‘K'”. This, too, is incorrect as it forgot Kenya.

In this regard, the company on Friday said the AI overviews for both queries have been removed for violating its policies.

In a statement, Google spokesperson Colette Garcia said that the vast majority of AI Overviews provide high-quality information, with links to dig deeper on the web.

Ms Garcia noted that several examples seen have been "uncommon queries", while the company has also seen examples that were "doctored or that we couldn't reproduce".

“We conducted extensive testing before launching this new experience, and as with other features we've launched in Search, we appreciate the feedback. We're taking swift action where appropriate under our content policies," the official noted.

At the bottom, every Google AI search overview acknowledges that the “generative AI is experimental.” The tech giant has stated that it conducts testing designed to imitate potential bad actors as part of its efforts to prevent false or low-quality results from coming up in the AI summaries.

These search overviews are part of Google's major plan to incorporate the Gemini AI technology across all of its products as it prepares to compete with rivals like Meta and OpenAI.

However, this was not the first time the company was forced to walk back the capabilities of its AI tools due to the issues faced by users. In February this year, Google had to pause its Gemini AI image generation feature after it offered “inaccuracies” in historical pictures.