Finland is making the most of artificial intelligence

Finns are tapping into the vast opportunities provided by artificial intelligence (AI). Sooner or later, intelligent machines will think, learn and react like human beings.

Artificial intelligence, a branch of computer science, can already perform demanding tasks, if taught and trained by humans.

In the future, intelligent machines will be able to learn like humans, act like humans, and think like humans. They can free us from tedious routine work, and will enable us to concentrate on more creative tasks that bring more value to our lives.

Three waves of AI

“The first wave of AI in the 1960s required coding and programming of rules, so that software and algorithms could solve specific problems,” says Harri Valpola, an accomplished computer scientist and CEO of The Curious AI Company.

“This enabled the creation of automated processes like route planning, which have become an integral part of today’s technology,” he continues.

“Today, when we talk about AI we refer to its second wave, which is based on supervised machine learning. Speech and image recognition, machine translation, data mining and other existing AI applications are all based on the second wave.”

Valpola says the third wave of AI, autonomous artificial intelligence, is emerging today. There are no third-wave technologies in current AI products yet, but research labs have had working prototypes for some time now.

It may take several decades before the intelligence of machines surpasses that of human beings.

“But things like digital coworkers that utilise a simpler form of AI will be around much sooner,” Valpola says.

Complex problem solving

Maria Ritola standing in a white spiral staircase.

“We are able to tap into knowledge that was never available to us before,” says Maria Ritola.Photo: Samuli Skantsi

“AI systems that identify patterns in vast amounts of data enable complex problem solving,” says Maria Ritola, the Finnish co-founder and CMO of Iris AI, which recently closed a two-million-euro funding round. “We are able to tap into knowledge that was never available to us before.” The startup has launched an AI-powered science R&D assistant that helps researchers track down relevant research papers without having to know the right keywords.

“But one of the risks of AI systems is that they learn human prejudices due to biases in the training data given to them, which is then used for decision making,” she says.

Social impacts of AI

“Another risk is that governments do not participate enough in developing AI systems,” says Ritola.

“As a result, we may fail to understand the social impacts of the machines that are getting ever more intelligent. One of the areas to understand and manage is the big shift in job markets relating to automation.”

Finland sees the big picture.

“The Finnish government is acutely aware that AI will change our jobs and careers, and wants to understand how it will affect individual people and our society,” says Pekka Ala-Pietilä, who heads a steering group that carved out a plan for Finland’s AI programme.

“Finland has huge potential to become one of the leading countries in exploiting the benefits of AI. The idea is to make it easy for businesses to utilise AI, and to support the public sector in building predictive, AI-powered digital services based on people’s major life events. We want to keep our country wealthy, our businesses competitive, our public sector effective, and our society well-functioning.”

AI MILESTONES

  • 1941
    German engineer and inventor Konrad Zuse builds the world’s first programmable and commercially available computer.
  • 1950
    British mathematician and logician Alan Turing introduces the Turing test, which lets people test whether a machine can think: The machine is intelligent if you can talk to it without noticing it is a machine.
  • 1956
    Researchers found a new academic discipline, AI research, at a workshop at Dartmouth College in the US.
  • 1961
    The first industrial robot, Unimate, starts work at the General Motors factory in New Jersey, USA.
  • 1982
    Finnish neural network pioneer Teuvo Kohonen introduces the concept of self-organising maps.
  • 1986
    American researchers Rumelhart, Hinton and Williams publish an article on MLP network and back-propagation, a new learning procedure that constitutes the basis for today’s deep learning AI.
  • 1997
    Chess computer Deep Blue beats the world’s best chess player, Garry Kasparov.
  • 2000
    Cynthia Breazeal of Massachusetts Institute of Technology in the US develops a robot called Kismet that can recognise and simulate emotions.
  • 2009
    Google starts to secretly develop autonomous, self-driving cars.
  • 2011
    Watson, a question-answering AI developed by IBM, can understand natural language. It competes against, and beats, two former winners of the quiz show Jeopardy.
  • 2012
    Deep learning technology beats all other computer vision methods in the ImageNet competition, where the goal is to recognise images in a vast set of approximately 1.2 million images.
  • 2012
    A robot that had learned to sort objects on its own, developed by Finnish robotics firm ZenRobotics, starts to sort useful waste material from industrial waste.
  • 2016
    AlphaGo, AI developed by Google, beats professional player and 18-time world champion Lee Sedol at Go, a complex game that requires creativity and is more difficult for a machine than chess.

By Leena Koskenlaakso, ThisisFINLAND Magazine 2018