Google Earth AI Tool Raises Concerns Over Misinformation

Google's new artificial intelligence tool, which allows users to generate synthetic satellite imagery for Google Earth, is facing scrutiny from experts who fear it could be exploited to create and spread misinformation. While Google states that its AI model incorporates guidelines to prevent the depiction of "harmful topics," investigations have revealed ways to bypass these restrictions.

BBC Verify demonstrated that minor adjustments to prompts could circumvent the AI's safeguards. For instance, a request to depict a "raised platform to hang traitors" near the UK parliament was initially rejected. However, a less specific prompt led to the image being generated. Similarly, a prompt to create the word "Trump" on the White House lawn was denied, but the tool allowed the creation of a community garden in the same location.

Experts express concern that the proliferation of AI-generated imagery, particularly in contexts historically perceived as untainted by artificial manipulation, erodes trust. "It's concerning because it's another place where sources of trusted information are being eroded by the constant doubt that AI might be involved," said one expert.

Google Earth serves as a primary entry point for many to access openly available satellite imagery, offering insights into remote or inaccessible areas. "This is potentially particularly damaging to public trust because satellite imagery has historically been seen as an especially trustworthy source of data," noted Bill Greer, a geospatial analyst and co-founder of Common Space. He added that falsified satellite imagery has already been a tool for spreading misinformation, and increasing governmental restrictions on data access may lead to a future with more fabricated images and fewer real ones, further diminishing public confidence.