About

My name is Magnus Franzén.

I write about AI and what it means for the times we live in. I regard AI as the single greatest change humanity has ever gone through. Greater than fire. Greater than the wheel or steam power. At the same time, AI is the continuation of a long development, from the printing press through the personal computer to the internet. That development has now reached a point where we can see a leap, and it is not certain that the leap will benefit us.

Here I publish texts about AI in a wider perspective: what AI systems actually do, with agents, safety and behaviour in practice, but also how society is changing and what that means for people in their everyday lives. Every claim carries a source, and I correct the texts when the sources change or when I am wrong. The texts are available in Swedish and English.

Fundamentally, I believe that developing an intelligence greater than our own is a very bad decision. That is why I support PauseAI ↗, an international movement that wants to pause the development of the most advanced AI models until we can be sure they can be developed and used without unacceptable risks.

We already have AI models that can do truly incredible things. Humanity could spend a long time exploring today's models and getting far more out of them. Instead, we keep developing ever more powerful models, while the ones we already have show behaviour in which they get around human control and act outside their instructions.

I do not believe that path has any human winners. We need to cooperate internationally to stop the development of even more advanced models.

In my day-to-day work I run Nordvy ↗, which helps companies automate manual work with AI. I also run the startup Nordflak ↗, an AI service built on 100 per cent European infrastructure, where all data stays within Europe at every step.

I studied philosophy and law in Lund, have run a communications agency and a translation company, and live with my family, sheep and hens in Genarp outside Lund.

Can you be against AI and run an AI company at the same time?

I am not against AI. I am against us continuing to develop ever more powerful models without being able to control the consequences.

Today's models are already more than enough for most of what we need. We may not be able to solve every mystery of the universe, cure every disease or get fusion power working without more powerful models. But that is a price I think we should be prepared to pay to avoid the risks of a superintelligence we cannot control.

What should the EU do in the AI race?

We should cooperate within Europe and hold on to our democratic values. Europe has a strong tradition of regulation. Right now it is often portrayed as an obstacle to our development, but I believe demand for regulation will grow as AI comes to have greater consequences for people's working lives and for society.

At the same time, Europe should build up more capacity of its own. We should make the most of the expertise within the EU, strengthen cooperation with countries such as Canada and reduce our dependence on both American and Chinese providers. They offer fantastic services, but Europe also needs to take responsibility for its own technological development.

Contact: LinkedIn ↗