European nations, Britain and Japan don’t need to make their own ChatGPT to compete in the artificial intelligence race, says a new paper from TBI.
Open Source: How Middle Powers Can Build Influence in the Age of AI argues that middle powers should instead make open-source AI a national priority.
The paper argues that states failing to support domestic open-source AI ecosystems will find themselves entering the age of AI at a significant disadvantage.
In the foreword to the paper, Ukraine’s Acting Minister of Digital Transformation Oleksandr Bornyakov says: “Ukraine has built and sustained one of the world’s most advanced digital states, even under wartime conditions, by relying on open-source software ... Openness isn’t optional. It’s survival.”
European nations and countries like Britain and Japan should make open-source AI a national priority to stay in the global race, according to a new paper from the Tony Blair Institute for Global Change.
Open Source: How Middle Powers Can Build Influence in the Age of AI, published today, argues that middle powers do not need to build their own DeepSeek- or ChatGPT-style frontier AI models to compete.
While the United States and China are likely to continue dominating frontier AI model development, this does not lock other countries out of capturing the economic, geopolitical, strategic and security benefits of AI.
Instead, the Institute argues that the real race will be won by how models are adapted, deployed and embedded across the economy and the state. Open-source AI – where the tech’s key building blocks are made openly available – will be mission-critical.
Rather than competing in frontier model development, the paper makes the case for building strong national and regional “open ecosystems” around AI. Middle powers must focus on the layers that turn models into real-world impact, such as open-source software, tools, data sets, benchmarks and standards, using these to drive adoption, innovation and state capacity.
While open models are increasingly capable, their real strategic value lies in enabling countries to adapt and deploy AI on their own terms. TBI argues that this will be central to countries’ future geopolitical standing.
As both the US and China increasingly promote open-source AI to diffuse their technologies globally, middle powers face the risk of new forms of lock-in through dependence on particular model families, software stacks or standards. The answer, the report argues, is not to reject openness, but to actively shape it by contributing to shared tools, benchmarks and data so that no single country or company sets the rules.
By investing in open tools, data and infrastructure, middle powers can capture value from AI without owning the most powerful models themselves.
The paper features a foreword from Ukraine’s Acting Minister of Digital Transformation Oleksandr Bornyakov, which illustrates how this strategy works in practice.
In his foreword, the minister said:
“Ukraine has built and sustained one of the world’s most advanced digital states, even under wartime conditions, by relying on open-source software, open data and open standards rather than proprietary systems or prestige model-building …
“Openness is the technical foundation of our cooperation model. Open source, open data and open standards are not nice-to-haves – they are the infrastructure of innovation in the AI era. When AI is used in defence, health care and government, you need to see inside it; open source makes that possible …
“Openness isn’t optional. It’s survival. It’s leadership.”
One of the paper’s most striking analyses demonstrates how much modern digital and defence infrastructure already depends on open-source software without appropriate recognition. Today, open-source code underpins an estimated 96 per cent of US civil and military codebases, yet open ecosystems remain largely invisible in many national AI strategies.
This “hidden infrastructure”, the paper argues, is where long-term advantage will actually be determined. The ability to inspect, adapt and secure AI systems, switch between models as technology evolves, and build specialised applications for health, energy, defence or public services all depend on open foundations.
Drawing on this experience, the paper sets out a practical agenda for governments, including establishing flagship open-source-AI programmes, treating open-source tooling as critical national infrastructure, curating strategic data sets, using public procurement to support open standards, and developing shared benchmarks to build trust and accelerate adoption.
Guy Ward-Jackson, the paper’s lead author, said:
“Chasing national frontier models is a distraction for most governments. Any nationally built frontier model would likely lag the global leaders within months, delivering little lasting advantage.
“This doesn’t mean these countries are doomed. The AI race will not be decided solely by who builds the biggest model, but also by who builds the strongest ecosystem around AI. For middle powers, open source is the most powerful lever they have – unlocking deployment without dependency.”