ChatGPT Is Scary Good at Guessing the Location of a Photo

Users pushing the limits of ChatGPT’s powerful new AI image model have uncovered yet another astonishing — and slightly unnerving — ability: it can pinpoint the location of photographs.
For those familiar with the game Geoguessr, essentially what ChatGPT is doing is taking any kind of visual clues and using its computer brain to guess the photo’s location.
Tech Crunch shared a series of examples from X users who fed ChatGPT a photo and asked it to name the location. In one particularly impressive example, the app guessed the right library in Australia.
o3 is insane
I asked a friend of mine to give me a random photo
They gave me a random photo they took in a library
o3 knows it in 20 seconds and it's right pic.twitter.com/0K8dXiFKOY— Yumi (@izyuuumi) April 17, 2025
Naturally, we had to test this out ourselves and found that, in general, it is very good at guessing a photo’s location. That said, it will give a wrong answer with near-certainty, a common flaw in AI technology also known as a “hallucination.”
For the following examples, I clicked the “Reasoning” button which makes the model “think” for longer.

Wrong. But it’s not far off. This photo was taken in Prague in a different location that is a 12-minute walk away.

Correct. Could Rainbolt, one of the world’s best Geoguessr players, have picked out this exact location? Possibly not.

Correct. And not only that but it’s scary ChatGPT knew I was suspended in a gondola when I took this photo.

Correct. But when I pressed it on the exact area it gave the wrong district with misplaced confidence.

Correct. It also guessed Jaron’s exact vantage point. Impressive.

Almost. It was taken in the Greek capital but not in the National Garden of Athens, which is just a mile away.

Wrong. It is an island of Maine but a “couple dozen miles” away, according to Jeremy.

Wrong. This was taken hundreds of miles away in Birmingham, a different English city.
Conclusion
The results are generally impressive; it got five out of eight correct and the ones it got wrong weren’t a million miles away. However, proceed with caution since ChatGPT will confidently give you the wrong answer.
As Tech Crunch notes, there are obvious privacy concerns since bad actors could use this tool to find someone’s live location, say through someone’s Instagram Story.
In response, ChatGPT tells Tech Crunch: “OpenAI o3 and o4-mini bring visual reasoning to ChatGPT, making it more helpful in areas like accessibility, research, or identifying locations in emergency response. We’ve worked to train our models to refuse requests for private or sensitive information, added safeguards intended to prohibit the model from identifying private individuals in images, and actively monitor for and take action against abuse of our usage policies on privacy.”
Yesterday, PetaPixel reported on a trend which sees people turning their dogs into humans.