July 31 (Reuters) - Alphabet's Google said on

Friday it was rolling back an AI image-generation feature it had

introduced in Google Earth just a day ago, after users shared

images that appeared to violate the company's policies.

The feature, powered by Google's Nano Banana 2 AI model,

had allowed users to generate photorealistic images grounded in

Google Earth's satellite, aerial and 3D imagery by typing a text

prompt for any location on the globe.

The tool had come under criticism for its potential for

misinformation as it allowed users to create fake scenarios on

top of real imagery of monuments and landmarks.

"We've seen geospatial professionals using this feature for

a range of useful purposes, however we've also seen people

sharing screenshots of generated imagery that appear to violate

our policies," Google said on X, adding that it was pausing the

capability while it works on stronger guardrails.

Google did not specify what types of images violated its

policies. It said the images did not appear in the main Google

Earth experience and were watermarked as AI generated.

While tech giants have rushed to integrate AI tools across

their products, they have had to adjust, restrict or temporarily

disable features in response to user violations, misinformation

and policy concerns.

Meta earlier this month discontinued Muse Image, an

AI feature that allowed users to generate images using public

Instagram accounts, after drawing widespread criticism over

privacy concerns, including from a Hollywood union.