In short
Google eliminated an AI image-generation characteristic from Google Earth on July 31, only a day after its July 30 launch, saying customers had been sharing generated imagery that appeared to violate its insurance policies.
Journalists and open-source researchers confirmed the Nano Banana software might simply fabricate occasions that by no means occurred—a blast crater in Los Angeles, a flooded U.S. Capitol, Iran’s Kharg Island on hearth—elevating fears it might supercharge misinformation.
Google’s protection that photos carry a SynthID watermark didn’t reassure critics, and the corporate mentioned it might solely restore the characteristic after including stronger guardrails, giving no timeline.
Google has pulled a newly launched synthetic intelligence characteristic from Google Earth barely a day after releasing it, following a swift backlash from journalists and open-source investigators who warned it might flood the web with convincing faux satellite tv for pc imagery.
The corporate launched the software on July 30, letting customers zoom to any location on Google Earth’s internet model, click on “create picture,” and generate a scene from a textual content immediate utilizing its Nano Banana mannequin. By July 31, it was gone.
In a press release posted to X, Google mentioned individuals “uniquely belief Google Earth for a dependable view of the world,” and that whereas geospatial professionals had discovered helpful functions, others had been sharing generated photos that appeared to violate its insurance policies. It mentioned it was rolling again the characteristic whereas constructing stronger guardrails.
The alarm centered on how simply the software fabricated occasions that by no means occurred. Tech outlet 404 Media demonstrated it might produce a blast crater in Los Angeles and add protesters exterior Google’s personal Mountain View campus. NPR generated photos of Iran’s Kharg Island ablaze and a flooded U.S. Capitol, each of which might be main information if actual. Open-source researcher Henk van Ess informed NPR he tried prompts together with refugees on the Mexican border and a nuclear plant in Iran, and that none had been refused.
Satellite tv for pc imagery has lengthy served as a trusted anchor for verifying breaking information and atrocities, exactly as a result of it has been tough to faux. Bellingcat researcher Jake Godin cautioned that one-click technology would streamline the creation of fakes and speed up their unfold, including that misinformation outruns any correction and that governments might now dismiss genuine photos as fabricated.
Google had initially downplayed the issues, noting that each picture carries its SynthID watermark, which flags it as AI-generated in instruments like Gemini, and that it blocks creation on dangerous matters. Critics referred to as that inadequate, arguing few individuals cease to confirm photos earlier than sharing them.
Google mentioned it might reinstate picture technology in Google Earth solely after implementing tighter protections, although it gave no timeline.
Day by day Debrief E-newsletter
Begin daily with the highest information tales proper now, plus authentic options, a podcast, movies and extra.
