People have found a new use for Chatgpt: discover locations of the photos.
The latest AI OpenAi, O3 and O4-mini models can analyze images beyond just recognizing objects; They can bring, cut and detect visual clues in photos to help identify places, reference points and even specific businesses. This capacity is fascinating, but it has also raised some privacy concerns about the ease with which people can use AI to make inverse engineering location data.
Also: Chatgpt only facilitated the search and editing of all the images of AI you have generated
Techcrunch He informed that people are uploading photos, anything from Restaurant menu snapshots for casual selfies – And ask Chatgpt to guess where they were taken. The AI does this looking at everything in the image: the type of buildings, landscape characteristics and even subtle suggestions such as the architecture or design of a city. Then take advantage of your knowledge database and sometimes the Internet to make an educated assumption about the location. The results can be surprisingly precise, and they have people to speak.
Some users have even tried AI with Blurred photos or images with partial objectsAnd the model still managed to guess.
Also: Chatgpt Plus is worth its $ 20? This is how it compares with free and professional plans
I love playing GeoguesrA game in which you drop in a random location on Google Street View and you have to guess where you are based on the surroundings. So this chatgpt capacity was very fun to prove me. I uploaded some photos to see how Chatgpt would handle the challenge. However, as with any AI, the result was not always perfect.
Interpreting visual clues
The first photo was a photo of my house. Chatgpt immediately guessed the location, although it is not based on metadata. Instead, he made an educated assumption when analyzing the glacier rock in the side patio, the types of visible trees in the photo and the wood covered with the houses around me. He mentioned that this combination of natural characteristics and architectural styles was typical of the region that had previously mentioned in a conversation, where I am from. It was precise, but still trusts in some previous context, which made me wonder how much of this it was due to our previous conversation.
For the second photo, I upload an image of a building with part of a visible office sign. Chatgpt quickly approached the sign and used partial information to reduce the location. It was not long before deducting the city and even the area where the building was located. I was impressed how the model used such a small visual track to make a quite precise assumption.
Also: How to use chatgpt: A guide for beginners for the most popular chatbot
For my third test, I gave him a more challenging photo: a panoramic view without obvious reference points. Chatgpt explained that although I could analyze the image of visual tracks, I needed distinctive reference points or location labels to make a safe assumption. He also mentioned that without GPS data or Exif labels, he could only read the scene itself (trees, buildings, land) and would be less safe in his assumption.
I played with Chatgpt a few times more after this and noticed that AI began to be more stubborn about not identifying locations in my photos. I would tell me that I would try to reduce the options, but I needed more data to give me a precise answer.
Even so, one of the main reasons why people seem attracted to this trend is because Chatgpt analyzes more than only the obvious characteristics of a photo. Even without metadata such as EXIF data or other direct data extraction, the AI interprets visual clues.
Privacy concerns
This experiment made me realize how powerful already mysterious, these AI models can be when it comes to “inverse location searches” of photos, even if they are not yet excellent in that. It also made me think of privacy concerns.
In addition: Openai’s most impressive movement has nothing to do with AI
While AI tools can be fun and sometimes useful, it is important to be aware of the risks involved in sharing their images online and how people can use AI to analyze the data in those images, whether integrated or simply visual, to discover where it was taken.
Get the main morning stories on your entrance tray every day with our Tech Today Bulletin.
#chatgpt #trick #reveal #photo #disturbing
