AI AND ML<\/p>\n
Turns out letting anyone rewrite the planet with a prompt had drawbacks<\/p>\n
Google performed a remarkably brisk reverse ferret last week, launching AI image generation in Google Earth and yanking it the next day after certain sections of the internet did the inevitable.<\/p>\n
This is how it began<\/a>: “Today we’re bringing AI image generation with Nano Banana to Google Earth, letting you virtually reimagine anywhere in the real world.”<\/p>\n This is how it’s going<\/a>, less than 48 hours later: “Thanks for your submissions, we appreciate it \u2013 unfortunately we have seen people sharing screenshots of generated imagery that appear to violate our policies. So we’ve rolled back this feature in Google Earth while we work on implementing stronger guardrails.”<\/p>\n Never mind “this is why we can’t have nice things.” What exactly did Google expect when it added AI image generation to Google Earth?<\/p>\n The process was simple \u2013 zoom into anywhere on Google Earth, hit “Create Image,” enter a prompt to twist satellite imagery to reflect your hopes and dreams, then enjoy your creation.<\/p>\n Or, as users quickly demonstrated, misinformation. While Google included cutesy examples of animal resorts, construction renderings, and restorations of ancient ruins, the darker side of the internet soon showed itself.<\/p>\n AI tools have long been able to alter screenshots, but placing the generator inside a service trusted as a record of the physical world does no favors for that trust, despite Google’s insistence<\/a>\u00a0that the output was watermarked as AI-generated.<\/p>\n In an update, Google stated: “We know that people uniquely trust Google Earth for a reliable view of the world,” which makes the functionality even more bizarre. As AI and misinformation expert Henk van Ess noted<\/a> in his Digital Digging blog, “What on earth is Google doing?”<\/p>\n The AI misinformation genie has long left its bottle when it comes to imagery. However, putting a photorealistic image generator inside Google Earth and expecting only harmless fun shows Microsoft Tay levels of naivete. In retrospect, Tay, the AI chatbot that netizens persuaded<\/a>\u00a0to spew some impressively hateful tweets, offered a warning back in 2016 of what was coming down the line.<\/p>\n