When AI development is unstoppable, what do we do?

In April 2026, Qwen 3.6 was released: an open-source AI model that anyone can run on their own hardware. It is capable of AI-assisted coding comparable to commercial, closed models (though it is not quite as advanced as the latest proprietary alternatives).

In 2023, Mustafa Suleyman, founder of DeepMind (the company behind AlphaGo) and current CEO of Microsoft AI, wrote in his book The Coming Wave: “From where we stand today, it appears that containing this wave is not possible.” It was an informed prediction. Now it looks as though we are living it.

Why won’t AI stop?

A joint report by Demos Helsinki, AI Finland and Business Finland shows that Finnish companies’ AI adoption grew by 13.5 percentage points in a single year. Philosopher Frank Martela wrote in Helsingin Sanomat that humanity’s greatest challenges, AI among them, demand agency, not merely hope. AI is no longer coming. It is already here.

Suleyman describes three structural forces in his book that make stopping AI development practically impossible. The first is capital: hundreds of billions of dollars flow globally into AI development and its applications each year, far too much for any single decision to reverse the direction (and the pace is only accelerating). The second is competition: if one country or company slows down, another gains ground, as we have now seen with China and the United States leaving the EU far behind. The third is dual-use potential: the same technology that improves medical diagnostics also develops autonomous weapons. Military interest does not disappear through regulation.

Will we fall behind or flourish?

Historian Yuval Noah Harari writes in his landmark work Sapiens that humans rose to dominate the world through one unique capacity: the ability to imagine things that do not exist, and to make others believe in them too. Gods, states, money, the rule of law: these are all collective products of the imagination upon which we have built our entire civilisation.

AI is the latest product of that same capacity. It is not a force of nature striking from outside, but the result of human imagination and engineering. Suleyman calls AI a “general-purpose technology”, in the same category as the steam engine, electricity, and the internet. Every time such a technology has arrived, some countries have fallen behind and others have flourished.

Finland has been a high-technology model country for decades, from Nokia to Linux, from Vaisala to Supercell. AI could be the next chapter of that same story. At the start of electrification, the countries that built the grid first gained a significant head start. We now face the same kind of choice.

What does accepting inevitability mean in practice?

Philosopher Frank Martela sums it up: “Agency must not be contingent on hope. Agency springs from serving a purpose greater than oneself. From having something worth fighting for.”

And separately on AI: “Developing AI that serves human flourishing rather than suppressing it is a worthy goal.”

Act first, and hope will follow.

Accepting inevitability is not surrendering. It is the starting point for agency. It is also essential to recognise that choosing not to decide is itself a decision: if we do not shape the development of AI and its applications, someone else will. The same applies to individual organisations: AI is not taking your job, but unmanaged change is a threat. My reading is that we have three concrete courses of action to pursue: active regulation, capturing the commercial benefits, and building digital independence.

How should Finland regulate AI?

The EU AI Act is now in force. It is not a burden but a tool, when used actively. Finland’s interest is to help shape its implementation and interpretation, not to wait and see how others do it.

The Nordic high-trust society and strong data protection tradition are competitive advantages, not obstacles: they lay the foundation for AI systems that both citizens and businesses can trust. These systems can also be exported to markets outside the EU. At its best, regulation is an export product, not merely a compliance obligation. Greater understanding of what the AI Act means in practice is needed, along with the capacity to participate as an equal partner in EU-level preparation.

How does Finland capture commercial value from AI?

Risto Siilasmaa has assessed publicly that the single biggest bottleneck preventing Finland from becoming a leading AI society is insufficient speed and scale of implementation, particularly in how strong capabilities translate into large-scale adoption, new companies, and attracting talent. We understand AI reasonably well, pilot it moderately, and regulate it, to some extent. But in commercialisation and scaling we fall behind: we produce expertise whose fruits are harvested elsewhere.

National computing infrastructure, meaning GPU capacity accessible over the long term to Finnish companies, researchers, and public services, would be an investment whose returns are measured in decades. The same applies to the skills base: AI capability is not built through courses, but through hands-on work and engineering. Strengthening the entire AI ecosystem requires that we not only buy foreign solutions but build our own capacity, so that the fruits of commercialisation stay here.

What does digital independence require from Finland?

At present, our options are open Chinese models or commercial American platforms. Neither is a neutral choice: both carry geopolitical risks and dependencies that only become visible in a crisis.

European cloud infrastructure and digital infrastructure based on Nordic cooperation are not merely ideological. They are an essential part of supply security. Nordic cooperation means in practice shared computing capacity, interoperable standards, and multilingual services that function without dependence on a single commercial operator. Control of our own data and the ability to deliver public services through domestic or European platforms is a concrete competitive advantage, including in situations where transatlantic trust is under strain.

Digital independence is built step by step. I wrote in March 2026 about the City of Helsinki’s example in my blog.

How do you think we should respond to the changes brought by AI?