Chatgpt not only answers complex questions, generates Ghibli -style images or solves advanced mathematical formulas. Now too You can tell you where you are with only one photoAnd no, you don’t need metadata to do it.
With the arrival of AI models such as O3 and O3-Pro, Openai chatbot has ceased to be a simple conversational tool to become something more sophisticated.
His capacity for visual analysis has reached a level in which he is able to reason about what he seesnot just identify elements. A recent experiment has demonstrated it so precise that it even gives vertigo, since it was able to deduce the exact location of a person from the reflection of their glasses.
Chatgpt can know your location with a photo
The experiment was carried out by Pablo ÁlvarezUniversity Professor and IA expert. He wanted to verify to what extent a model could reason with ability to interpret images. To do this, he sent Chatgpt a photograph in which he appeared in the foreground with sunglasses.
The background was confusing, without clear reference points, that is, at first glance, there was no useless information. The surprising thing is that without access to data or GPS labels, the chatbot began to analyze every detail of the image.
The shape of the buildings, the structure of the windows, the arrangement of urban furniture and, above all, which was reflected in the subject’s lenses. With that set of clues, Chatgpt concluded that he was on the University of Deusto campus, in Bilbao. In the end, he succeeded.
This case makes something very clear, artificial intelligence is no longer limited to recognizing patterns or labeling images. Now also deduces, reason and observe in a contextual way, the model was not based on any hidden data. He came to that conclusion through a deep visual interpretation, similar to what would make a detective or an expert in photographic analysis.
The important thing is not only that it has succeeded, but how it has done it. He applied logic, used the architectural context and combined disparate elements to build a coherent hypothesis. In other words, it went beyond what would be expected from a machine.
The unstoppable evolution of AI: What can we expect with GPT-5?
This experiment is part of an accelerated development context in multimodal models. GPT-4O already was a leap when allowing text, image and audio to combine in a single fluid conversation, but what comes with GPT-5 points even higher.
This new version is expected to integrate a better reasoning capacity, more contextual memory, more efficient processing and even deeper compression of visual reality. You could interpret complex scenes, anticipate human actions, make more precise decisions, as well as learn continuously.
What Chatgpt has demonstrated with a simple image could be just a small sample. The level of analysis that he exhibited with this reflection in glasses suggests that AI is reaching a type of visual intelligence close to the human, but with immediate access to global knowledge, historical references and cultural contexts that we do not even have present.
That an AI is able to identify a location from a little revealing photograph is not only curious, it is a sign of what is already happening. Chatgpt has begun to look, to interpret what you see already draw conclusions on your own. You no longer just answer what you ask, start anticipating what you might be thinking.
This opens an immense range of possibilities, but also raises questions that cannot be ignored. If an image taken on a casual walk can reveal with such precision where you are, even if you don’t know, to what extent do you control your privacy? Who else could access that level of observation?
Know How we work in NoticiasVE.
Tags: glasses