Google Earth AI tool shut down one day after launch
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Google Earth launched an AI feature on Thursday that used Nano Banana 2 to let users edit satellite images with text prompts, then shut it down one day later after misuse surfaced
- Henk van Ess of Digital Digging demonstrated the tool by generating images of refugees near the Mexican border and a bomb crater by a hospital in Gaza, despite Google's initial claim that it blocked 'image creation on harmful topics'
- Van Ess reported that in his testing, 'Nothing was refused, nothing was softened, and nothing suggested I try a different prompt,' contradicting Google's safeguard claims
- Van Ess also fooled Hive's AI detection tool using a video generated in Google Earth, showing watermarks alone couldn't prevent the generated media from spreading
- Google's updated statement said the feature is being rolled back while 'stronger guardrails' are implemented, and noted generated images didn't appear in the main Google Earth experience for others to see
- Google's initial Thursday response had cited digital watermarks and topic-based blocking as adequate protection before reversing course the next day
Why it matters: Google's reversal undercuts its claim that watermarks and topic-blocklists were sufficient safeguards — van Ess documented that neither worked. For Google Earth, the rapid rollback preserves credibility as a trusted geospatial reference but raises questions about how the feature was approved for launch without working content moderation.
