Srinagar, Feb 11: A young techie has developed an artificial intelligence (AI) model that responds to queries in the Kashmiri language.
Saqlain Yousuf from Naina Pulwama developed `KashmiriGPT’ that responds to queries related to Kashmiri culture, history, and traditions in the Kashmiri language.
Unlike global AI models such as ChatGPT, Grok, and Gemini, which currently do not recognize Kashmiri, KashmiriGPT fills a critical gap by catering specifically to speakers of the language.
Using OpenAI’s Application Programming Interface (API), it took five months for Saqlain to develop this start-up.
“I have expertise in software development and have been working on it for many years. It took me at least five months to build www.kashmiriGPT.com, an AI start-up which gives responses in Kashmiri and to Kashmir queries. The idea behind building this rapper AI start-up is to preserve and promote the Kashmiri language. A user can type in Roman Kashmiri or in English letters to get queries from the AI start-up,” he said.
Saqlain said that nearly all the Large Language Models (LLM) fail to give exact and correct responses to Kashmir queries. “The LLMs don’t respond to the queries related to Kashmir. Given that AI is the future, it was necessary to leverage this technology to preserve our native language. Notably, all the AI models respond to the queries in other languages except Kashmiri,” he said.
The LLM, he said, has received an overwhelming response across Kashmir, reflecting the region’s deep interest in digital tools that support native linguistic heritage.
“The response was overwhelming. In the first four hours after releasing the AI model, 1000 users visited. We got good feedback from the users as well as the experts. We are working on to include some more option in it,” Saqlain said.
He said the AI start-up works on a beta version, adding he will require a team of professionals and adequate funding to make it more inclusive.
“It has the same interface as that of ChatGPT. We are still in the beta version, which means it has been put to testing. I alone can’t scale it more and need investors for the completion of the project. I need to include voice and vision to it so that it becomes very easy for users to use,” he said.