Debasis Mohapatra
Bengaluru, 30 March 2025
AI (artificial intelligence) models like ChatGPT, Grok, Gemini, DeepSeek & Anthropic among others are bringing sweeping changes in the technology world. All facets of software development are getting redefined with these LLM (Large Language Models). Leveraging these AI tools, many sectors are increasing efficiency significantly. For instance, coding time is getting shorter; content development has become very easy; automation is making many repetitive jobs, free of human intervention.
While many such changes are likely futher accelerate in coming days with rapid progress in AI field, one question comes to our mind- will it remain open source? First, let’s understand the meaning of open source AI model.
In an open source AI model; code, architecture, and even trained parameters (weights) are made freely available to the public under an open-source license. This means anyone can access, use, modify, and distribute it, often with minimal restrictions, as long as they follow the license terms.
As per this definition, open source models are not many. For instance, ChatGPT, developed by OpenAI, is not open-source. It means its code, underlying model are not available for modification or redistibution. OpenAI keeps strict control over its proprietary technology.
In comparison, xAI’s Grok-1 Model is an open-source AI model. But its later releases like Grok-2, and Grok-3 are not open source AI models. Similarly, Google Gemini is considered to be a closed model. Meanwhile, China-origin DeepSeek model is a open source one with all its latest models being available to developers.
More closed models releases:
Beyond initial releases, promoters of most AI models are now preferring their tools to remain closed ones. This trend is driven by multiple factors. Firstly, owners of these large LLMs have pumped billions of dollars into building these models. Now, they want return on their investments. That is the reason that higher versions of most LLMs remain proprietary. It means the technology is not shared with anyone. Secondly, there is a geopolitical angle to AI race now with US & China fighting it out. US, which has a clear lead, wants to dominate the scene in this critical technology area. Given these factors, future models are likely to be closed ones.
What did Indian IT minister say?
Recently, Indian IT minister, Ashiwini Vaishnaw said that the open source models may cease to exist in future. And that is what driving India to build its own foundational model. According to the minister, India will be able to build its own LLM in the next 10 months, leveraging open source models. Not only that, the country eyes to have its own high-end computing chip sets- GPUs, in the next 3-5 years. There will be many nations, which will follow the path of India as AI becomes a critical part of national security.
