1887

Experimental Finland

Written by: Clara Young, OECD Observer
Last update: 29 January 2020

Start-up event, Helsinki, Finland

Most people spend their evenings kicking back with a book or whatever’s on Netflix. In Finland, they’re teaching themselves artificial intelligence (AI). In the summer of 2017, computer scientist Teemu Roos heard that the government was looking for ways to teach ordinary people the basics of artificial intelligence. It would be a continuing education initiative-—not necessarily to train people to become machine-learning engineers, but to understand how neural networks work and grasp how AI is changing the way we do things.

Roos teamed up with the tech firm Reaktor to design an easy-to-follow massive online course that rolled out in May 2018. Some 90,000 people signed up. By the following September, 7,500 people had completed the 30-hour course and graduated. The goal the Finnish government has set is to educate 1% of its population about machine learning.

“Elements of AI” is just one of Finland’s many pilot projects and social experiments, big and small. This is something the Finns are really good at: carefully designing long-running, society-wide experiments with broad grassroots participation that test ideas for making society run better. The radical idea was to turn this penchant into a national experimental culture.

What the government has called “Experimental Finland” came into being in 2015 in the prime minister’s office. The goal was to set up small “sand-box” projects and larger formalised policy trials in the areas of circular economy, digital workforce skill and artificial intelligence over the next 10 years. A small Experimental Finland team was put together to call for new ideas and oversee experiments at all different levels of government, from municipal to national, and across the country. Projects and trials that pass muster are funded by the government or co-financed by the public and private sector. Finland has taken a systems design approach in its policymaking. And it’s telling its civil servants not to be afraid to try out new ideas.

Finland’s complicated social security, for instance, is a system in need of new ideas.

Finland started out with a small residence-based pensions programme and universal child benefits after World War II. Today, it provides mandatory day-care services for families with small children and home-care benefits for those looking after members of the family who are ill or elderly. But Finland’s welfare state needs streamlining. And the current system of tying people’s social protection to their jobs is rapidly growing out-dated: people are turning to temporary jobs, increasingly working as independent contractors, or, perhaps as a fall-out of automation, not able to find work at all. Like many other countries in the world, Finland thinks universal basic income (UBI) is something worth looking into, and now it has.

Preparation for a formalised trial on universal basic income began in 2015. Among the more innovative ways the government elicited research on basic income models was to organise a two-day Basic Income Hack. For 32 hours straight, 10 teams composed of coders, researchers, politicians, communications specialists, graphic designers, activists and information designers brainstormed about basic income. Some of their ideas, including a basic income game that simulates how certain life choices affect public finance, shaped the design of the UBI trial.

On 1 January 2017, the two-year trial began. The Finnish social security agency, Kela, paid a randomly sampled, non-voluntary group of 2,000 unemployed people aged between 25 and 58 a monthly basic income of €560, unconditionally and without means-testing. This replaced their usual unemployment benefits. The control group was also unemployed but received their benefits in the traditional way.

The experiment ended on schedule on 31 December 2018. The following February, Kela released its preliminary findings: “The basic income experiment did not increase the employment level of the participants in the first year of the experiment. However, at the end of the experiment the recipients of a basic income perceived their wellbeing as being better than did those in the control group.” Based so far only on data from the first year of the trial, there was no difference between the group that received basic income and the control group in terms of finding work. Both groups also worked an average of 49 days during that year, with the UBI group earning €21 less than the control group. Findings from the second year of the trial will be published in 2020.

To encourage more experiments of this kind, the government has opened a digital platform called kokeilunpaikka. It means Place of Experiment. Here, people can read up on results and analyses of past experiments, find out what sorts of projects the government is looking for, and go through a step-by-step process of submitting their own experiment. In healthcare innovation, for example, a group of nurses is testing robotic vests that may help them lift up patients more easily. In another project, a game is being piloted that helps prepare medical patients for diagnostic testing. Sometimes experiments can run into snags because of legislative obstacles. The government hopes to smooth these out by bringing out a guide that helps navigate through these situations.

Finland is always ready to take something new out for a spin. Where else could a pulp and paper company have come up with what was, for a time, the world’s most coveted cell phone? But innovation is typically a private sector asset. Turning it into a public sector one too is perhaps the grandest experiment of all.

©OECD Observer March 2019

References

www.kela.fi/web/en/basic-income-experiment-2017-2018

OECD (2017), OECD Reviews of Innovation Policy: Finland 2017, OECD Reviews of Innovation Policy, OECD Publishing, Paris, https://doi.org/10.1787/9789264276369-en. https://doi.org/

www.oecdobserver.org/finland50oecd

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error