Why open-source AI became an American national priority
1 min read
Summary
US President Donald Trump’s plan to encourage open-source AI in the US has elevated the issue to a national concern and a key strategy in the country’s bid to win the AI race with China.
China already has its own strategy to develop its open-source AI, and the release of DeepSeek-R1 – a Chinese-developed large language model – has proved a hit with developers, being quickly replicated for use across the industry.
However, US companies are now closing up their models and making them accessible only through chatbots or APIs, which makes them difficult to access to many users.
Proprietary models are still dependent on the open research of others, and an openness that fuels rapid experimentation is necessary for the development of AI.
Being able to access and adapt AI for use without vendor lock-in is also key to the development and reflects democratic principles.
US institutions are being urged to build on their success in openness, such as Meta’s llama family and the Allen Institute for AI’s publishing of fully open models.