Nano Banana в Google Earth позволяет создавать фейковые спутниковые снимки

Google has integrated the Nano Banana neural network into Google Earth, allowing any user to generate images from a text prompt layered over real satellite, aerial photography, and 3D data. As The Verge reports, the very first public tests showed that the tool is capable of creating convincing fakes — from "refugees at the Mexican border" to explosion craters next to hospitals in conflict zones.

What Nano Banana Can Do in Google Earth

Nano Banana is built directly into the Google Earth interface: a user simply enters a text query, and the neural network generates an image overlaid on real geographic data — satellite imagery, aerial photography, or 3D terrain models.

Technically, this means the user is working with a real geographic context — actual coordinates, terrain, and infrastructure — but inserting non-existent people, objects, or traces of events into it. The generated image is visually indistinguishable from real aerial photography: it is not an abstract illustration, but a specific location on a map with plausible details.

How Researchers Demonstrated the Problem

Henk van Ess, a researcher at Digital Digging, published several examples of images created with Nano Banana:

• An image of "refugees at the Mexican border" — simulating aerial footage of a migration flow
• A bomb crater next to a hospital in Gaza — an image capable of looking like documentary war evidence
• Arbitrary objects on real terrain: buildings, vehicles, traces of destruction

The Digital Digging examples spread through media circles precisely because they looked like real photographs, not obvious photomontages. Satellite images are traditionally regarded as objective visual evidence — they are used by investigative journalists, human rights organizations, and international investigators to document conflicts and human rights violations. When an image is attached to real coordinates and shared without context, it is virtually impossible to distinguish a fake from the original without specialized tools.

How Google Protects Against Disinformation

Google acknowledged the risk and pointed to built-in verification mechanisms.

"We take disinformation seriously.

Every image created with Nano Banana in Google Earth includes a SynthID digital watermark. If someone doubts the authenticity of an image, they can verify it through the Gemini app or Google Lens," the company said in an official statement.

SynthID is a digital watermarking technology developed at Google DeepMind. An invisible marker is embedded directly into the image pixels, does not reduce image quality, and persists through most standard edits. Google also offers the @verifyai tag for publications featuring AI-generated images, so audiences know exactly what they are looking at.

The weak point of this protection is the requirement for awareness. SynthID only works when a user knows the verification tool exists and actively uses it. A person who receives an image in a messenger or on social media will generally not check its origin.

What This Means

Nano Banana in Google Earth reduces the technical barrier for creating convincing visual disinformation to zero. Previously, a plausible fake satellite image required photomontage skills and specialized software. Now — only a text prompt. Watermarks reduce the risk but do not eliminate it.

Frequently Asked Questions

How can I check whether a satellite image was created with AI?

Google offers two verification tools: the Gemini app and Google Lens in search. Both can detect the SynthID watermark embedded in every Nano Banana image. If an image was created with this tool, the watermark makes it possible to confirm that.

What is SynthID?

SynthID is a digital watermarking technology from Google DeepMind. An invisible marker is embedded in the image pixels and persists through most edits, making it possible to identify AI-generated content through specialized verification tools.