A Goa-based cybersecurity professional has found an unconventional use for ChatGPT, using the artificial intelligence chatbot to locate his missing footwear among rows of shoes outside the Mahakaleshwar Temple in Ujjain, Madhya Pradesh.
Shubhang Borkar documented the experiment in a video titled "Used ChatGPT to find my lost clogs", showing how he used his smartphone camera and ChatGPT's image-analysis capabilities after struggling to locate his footwear outside the crowded temple.
Borkar, who describes himself on social media as a cybersecurity professional during the week and an explorer on weekends, initially recorded the large collection of footwear around the temple area. With several similar-looking pairs placed together, visually identifying his own clogs had become difficult.
He then opened ChatGPT and uploaded photographs showing different sections of the footwear area. Borkar also provided the chatbot with a reference image indicating what his missing clogs looked like.
He informed ChatGPT that he was at the Mahakaleshwar Temple and had lost his footwear among the shoes left outside. He then asked the AI system to analyse the images and help identify a potential match.
The process did not rely on a single photograph. As the search continued, Borkar uploaded images of different sections containing footwear, effectively giving the chatbot additional visual information to analyse.
ChatGPT eventually pointed him towards a specific pair visible in the photographs. Borkar then followed the suggestion and was able to locate his missing clogs, according to the video documenting the experiment.
While the incident represents a relatively simple use of artificial intelligence, it demonstrates how multimodal AI tools are increasingly being used for tasks involving both visual and conversational inputs.
ChatGPT's multimodal capabilities allow users to submit images alongside text-based instructions, enabling the system to analyse visual information and respond to questions about what appears in an image. Such capabilities have expanded the role of generative AI beyond conventional text generation and question-answering.
Visual AI systems can be used for tasks ranging from interpreting documents and analysing photographs to identifying objects and extracting information from images. In Borkar's case, the same underlying capability was applied to a personal, everyday search problem.
The incident has also gained attention on social media because of the unusual context in which the technology was deployed. Rather than using AI for coding, research, writing or workplace productivity, Borkar effectively turned ChatGPT into a visual assistant for identifying a physical object in a crowded environment.
However, the experiment should not be interpreted as demonstrating a dedicated lost-object tracking feature within ChatGPT. The chatbot was analysing photographs supplied by Borkar and suggesting a visual match based on the information available in those images.
The episode offers a small but practical example of how users are experimenting with multimodal AI beyond established professional applications, applying image recognition and conversational interfaces to routine situations encountered in everyday life.