The three-day AI Conclave featured a 'National Level AI Summit', project exhibition, robotics competition, e-games, tech talks and expo, panel discussion on 'AI for Society', discussions on AI policy, and public awareness programs related to AI.
We use Google Cloud Translation Services. Google requires we provide the following disclaimer relating to use of this service:
This service may contain translations powered by Google. Google disclaims all warranties related to the translations, expressed or implied, including any warranties of accuracy, reliability, and any implied warranties of merchantability, fitness for a particular purpose, and noninfringement.
The 'AI Conclave 2026' organized by the Kathmandu University Artificial Intelligence Club (KUAIC), under the Department of Artificial Intelligence of Kathmandu University in Dhulikhel, has concluded.
The AI Conclave, which was held for three days on January 12, 13 and 14, included a ‘National Level AI Summit’, project exhibition, robotics competition, e-games, tech talks and expo, panel discussion on ‘AI for Society’, discussion on AI policy, and public awareness programs on AI.
Prof. Dr. Sudan Jha, who was the keynote speaker of the AI Conclave, said that the artificial intelligence (AI) technology being developed globally is heading towards a serious intellectual crisis and Nepal should focus on developing ‘lean AI’ according to its own needs.
According to him, as the original human content on the internet is decreasing, AI has come to a point where it has to repeat itself by reading AI’s own creations, which experts have called ‘model collapse’ . ‘We have considered big AI technologies like ChatGPT to be everything,’ he said, ‘but the reality is different, as AI depends on its own content, its creativity and reliability are weakening.’
Prof. Dr. Jha said that in such a situation, a new wave of ‘small language models’ has started in the global market, creating an important opportunity for developing countries like Nepal. He said that this has opened up the possibility of developing small but effective AI models based on local language, culture and data.
According to him, the competition to make large, expensive and highly energy-consuming AIs around the world is also causing energy and water crises, Prof. Dr. Jha clarified that such ‘clumsy’ AI is not suitable for Nepal. ‘We need ‘lean AI’ that is low investment and runs on low electricity, but effective in local problems,’ he said.
According to him, chatbots can create literature, but small but accurate AI models will be more useful in tasks such as crop disease identification, herb collection, health service support and agricultural monitoring in remote villages.
Prof. Dr. Jha said that the current AI technology is highly dependent on the cloud, which causes problems such as Nepali data going to foreign servers, making services expensive and slow. To avoid this, Nepal should not take the risk of 'GPU debt' and move towards 'edge AI' and small models that run on local devices. He said that the 'small language model' can be operated on mobile or local systems even without the internet, so it can give immediate results in hydropower projects, health posts and services in remote areas.
Concluding that the future does not lie in big and clumsy AI but in 'lean AI' that is suitable for Nepali soil, balanced and smart, he suggested that AI should no longer imitate others, but learn from the original thinking, knowledge and actions of Nepalis.
