- From Singapore to Switzerland, many countries are developing indigenous AI language models to serve domestic needs. In Singapore, SEA-LION can speak 11 Southeast Asian languages, while Malaysia’s ILMUchat understands the difference between “Georgetown” in Penang and Georgetown University (US). Switzerland also launched Apertus, an AI model that understands the proper use of “ss” and “ß” in Swiss German.
- This is part of the “sovereign AI” trend – an effort by countries like the UK, India, and Canada to independently develop AI technology instead of relying on the US or China.
- However, according to expert Trisha Ray (Atlantic Council), developing an AI model from scratch is a “colossal burden” for mid-sized countries.
- India has spent $1.25 billion on the IndiaAI Mission program, supporting the company Soket AI in building a national AI model. The main reasons: data security and national defense sovereignty. Indian agencies are hesitant to use US or Chinese models for fear of data collection or “sovereignty distortion” such as DeepSeek potentially claiming Ladakh does not belong to India.
- Singapore chose a different approach: developing indigenous AI to “understand regional culture and language,” while complementing rather than replacing large models like ChatGPT or Gemini.
- A new proposal from Cambridge University called “Airbus for AI” calls for middle-tier countries – the UK, Germany, Canada, Japan, Singapore, etc. – to form an alliance to develop public AI to compete with US-Chinese giants.
- However, many experts, such as Tzu Kit Chan (Malaysia), warn that “the pace of AI technological development is currently too fast,” making national projects easily consume billions of dollars in vain. He believes that instead of competing on models, governments should invest in legal frameworks and AI safety.
📌 With over $1.25 billion from India and hundreds of billions globally pouring into “sovereign AI,” many nations hope to escape US-China dependence. However, experts warn that most users still choose ChatGPT or Gemini, risking national projects becoming expensive but futile races if they lack a collaborative vision and realistic usage strategy.
