Artificial intelligence has now gone beyond being an emerging technology; it has become an integral part of everyday life as well as industrial and scientific infrastructures. But in this fast-paced journey, how can the Persian language play an effective role? Are global AI technologies sufficient on their own, or do we need to develop and strengthen localized models tailored to our culture and linguistic structure?
Part Artificial Intelligence Research Center, with nearly seven years of experience in natural language processing, organized a specialized conference titled "Artificial Intelligence and Persian Sugar; On the Path of Part’s Innovations" to address these questions. This event provided an opportunity for researchers, policymakers, and AI industry professionals to come together and discuss the challenges, achievements, and future outlook of developing localized technologies for the Persian language. Amir Mohammad Saleh-ouf, head of Part’s NLP team, in a detailed interview, explained the goals and accomplishments of the event, emphasizing the importance of technology localization and the necessity of establishing a domestic AI ecosystem. Below, you will find more details from this conversation and the outlook for Persian language technology.
Strong Foundation and Movement Towards Local Innovation
Saleh-ouf explained that since 2017 (1396 in the Iranian calendar), the Part Research Center has been continuously working on developing natural language processing models and language models. He said, “During this time, our team has experienced significant growth and successfully completed numerous projects. The emergence of large-scale language models globally, which are now highly advanced and popular, motivated us to present the results of our efforts in a specialized event. Our main goal, besides reporting on our performance, was to create a platform for interaction and exchange of ideas with other AI professionals to better shape the future path of this technology.”
Building an Independent Ecosystem: A National and Regional Challenge
He spoke about the importance of such events in developing Iran’s AI ecosystem: “Unfortunately, the focus in our country has mostly been on using foreign products, with less attention paid to developing indigenous technologies. Meanwhile, we believe that building the fundamental infrastructure for artificial intelligence must come from within the country itself, relying on domestic capabilities. Events like ‘Artificial Intelligence and Persian Sugar’ play an important role in boosting motivation, self-confidence, and solidarity among local experts, and demonstrate that the Persian language requires dedicated models and solutions to best address its unique challenges.”
Localization: The Key to Preserving Linguistic and Cultural Identity
Saleh-ouf also pointed out a very important issue: “If artificial intelligence is trained solely on non-native language data, it cannot fully understand and convey all the nuances, cultures, and unique meanings of the Persian language. Therefore, developing native models is not only a technical necessity but also a cultural imperative that helps preserve our language’s identity and cultural values.”
Future Outlook and the Need for Convergence
The researcher emphasized that for genuine progress in the field of native artificial intelligence, continuous collaboration between universities, industry, and policymaking bodies is essential. He said, “Conferences like this event provide a golden opportunity to foster such convergence and can serve as a launching pad for innovative and practical projects.”
