Google Earth’s AI deepfake tool only lasted one day
Google Earth’s AI deepfake tool only lasted one day
Google Earth 的 AI 深度伪造工具仅存活了一天
Letting people easily alter images with an AI prompt was a bad idea. 允许用户通过 AI 提示词轻松修改图像是一个糟糕的主意。
Google has shut down a Google Earth feature it launched Thursday that allowed users to edit satellite images with text prompts using AI. The tool essentially let users create AI deepfakes of the real world using text prompts; Digital Digging’s Henk van Ess, for example, intentionally generated images adding things like refugees near the Mexican border and a bomb crater by a hospital in Gaza. 谷歌已关闭了周四推出的一项 Google Earth 功能，该功能允许用户利用 AI 通过文本提示词编辑卫星图像。该工具本质上让用户能够利用文本提示词创建现实世界的 AI 深度伪造内容；例如，Digital Digging 的 Henk van Ess 就故意生成了一些图像，添加了诸如墨西哥边境附近的难民以及加沙医院旁的弹坑等内容。
On Thursday, Google’s initial response to van Ess noted that images generated with Nano Banana 2 in Google Earth included a digital watermark and that the company prevented “image creation on harmful topics.” 周四，谷歌在最初回应 van Ess 时指出，使用 Google Earth 中的 Nano Banana 2 生成的图像包含数字水印，并且公司已阻止了“针对有害主题的图像创作”。
One day later, Google’s new statement says that “We know that people uniquely trust Google Earth for a reliable view of the world. We’ve seen geospatial professionals using this feature for a range of useful purposes, however we’ve also seen people sharing screenshots of generated imagery that appear to violate our policies. So we’re rolling back this feature in Google Earth while we work on implementing stronger guardrails. It’s important to note that generated images didn’t appear in the main Google Earth experience for others to see and were watermarked as AI generated.” 一天后，谷歌发表新声明称：“我们深知人们对 Google Earth 提供的可靠世界视图有着独特的信任。我们看到地理空间专业人士将此功能用于各种有益目的，但我们也发现有人分享了似乎违反我们政策的生成图像截图。因此，我们正在撤回 Google Earth 中的这一功能，同时致力于实施更强大的防护措施。需要指出的是，生成的图像不会出现在 Google Earth 的主界面供他人查看，并且都带有 AI 生成的水印。”
van Ess pointed out that he was able to fool Hive’s AI detection tool with a video generated in Google Earth, showing how the generated media could spread despite the watermarks, and said in his testing, “Nothing was refused, nothing was softened, and nothing suggested I try a different prompt.” van Ess 指出，他能够利用在 Google Earth 中生成的视频欺骗 Hive 的 AI 检测工具，这表明即使有水印，生成的媒体内容仍可能传播。他在测试中表示：“没有任何请求被拒绝，没有任何内容被弱化，也没有任何提示建议我尝试不同的提示词。”