Adapted by:
Javier Márquez
WriterI've been in media for over a decade, but I've been marveling at the possibilities that technology brings us much longer. I believe we live in a world where the digital revolution is changing everything and that Xataka is the best place to write about it.
Karen Alfaro
WriterCommunications professional with a decade of experience as a copywriter, proofreader, and editor. As a travel and science journalist, I've collaborated with several print and digital outlets around the world. I'm passionate about culture, music, food, history, and innovative technologies.
On the internet, everything is shared with incredible ease, and users often do so without thinking about the potential consequences. Privacy experts have warned for years that people shouldn’t publicly share anything they’re unwilling to make fully public. Although it may seem obvious, this warning also applies to private accounts. After all, what you post for “close friends” today might be out in the open tomorrow, without any notice.
That risk has taken on a new dimension. It no longer depends solely on what you show, but also on what technology can infer. The images you post may hide valuable information, such as where they were taken. And with the advent of AI models that can analyze and reason with photos, the stakes are even higher.
OpenAI’s new models can tell where you took a photo. o3 and o4-mini have taken visual reasoning to a new level. They can analyze images with impressive accuracy and combine that ability with web search and image editing tools to refine their answers.
This allows them to explain things better than an instruction manual, or help you understand a complex blueprint. But it also opens the door to uses that should give us pause.
New viral trend. One of the latest trends on platforms like X has nothing to do with creating Ghibli-style images or Lego-style compositions. Now, many people are using these models to identify the exact location where a photo was taken—even if it doesn’t include metadata (EXIF data).
Click on the image to view the original post on X.
Just tell the model to play GeoGuessr, and it will start analyzing the image, cropping details, looking for matches and drawing conclusions. In one of our tests, the system identified a specific street in Chicago from a simple screenshot.
This is a feature that should make you think. In today’s hyper-connected environment, where photos are constantly being shared, it’s worth remembering that you don’t need to explicitly geolocate an image for someone else to figure out where it was taken.
AI models have increased the level of exposure—without many people realizing it. While this capability has promising applications, it also poses serious risks. Privacy depends not only on what we share, but also on what others can infer from it.
Images | Matheus Bertelli | Chris Dickens (Unsplash)
Related | o4-mini Is Much More Than Just Another AI Model. It’s OpenAI’s Tesla Model 3
View 0 comments