Will we get the artificial intelligence we deserve?

We have picked a challenging theme this time: artificial intelligence. Everyone agrees this will have major consequences across nearly all areas, but nobody can predict exactly what might happen. 

We are facing a situation similar to that in the 1990s when the internet was introduced. 

“Nobody knows what will happen,” acknowledges Svein Berg, Managing Director of Nordic Innovation.

“But what we do know is that it will happen fast. As an American researcher put it in the 1960s: The pace of change has never been this fast, yet it will never be this slow again.

We have indeed been thinking about the consequences of AI more or less since the emergence of what is considered to be the first computer, Eniac, in 1947. 

Hannes Alfvén, who was awarded the Nobel Prize in Physics in 1970, some years earlier published a story originally meant for his children and grandchildren: The Tale of the Big Computer. 

In it, a computer tells the story of how the age of data began. Humans were only a step on the way to the high point of evolution: the computers.

The problem with the human was that she was thinking too slowly. On their own, humans could predict what the weather would be in three days. But it took three days to make that prediction, so it was already outdated by the time it was finished. Without the help of computers, it was not possible to predict the weather.

Of course, you could imagine many humans working together, effectively linking their brains. But it did not work since their way of communicating – through numbers or in writing – was too slow and imprecise, the computer in the story pointed out.

In field after field, computer programs have been developed that are faster than humans. When IBM’s Big Blue computer finally beat the reigning world chess champion Garry Kasparov in 1997, it was considered to be a milestone in the development of artificial intelligence. Then followed programs that made computers superior in everything from Go to Jeopardy.

Nothing stops these programs from being linked in series to create a computer that can beat humans in all games. But can you link enough programs to make a computer better at all intellectual tasks that humans can perform? If so, this would lead to the emergence of what is known as Artificial General Intelligence (AGI). 

With eight billion people who constantly invent new things, this probably cannot happen until computers start building their own computers. The possibilities are almost limitless. In this edition, we look at some aspects of AI:

• Autonomous ferries
• Drones repairing wind turbines
• Programs that diagnose illnesses
• Programs that help us in the office
• The opportunity for Nordic data servers

Common for all these is that the AI system works quicker than humans can, but things might go wrong if we do not know how to handle the programs. That is why education is so important. In Finland, an AI course has been developed that has become an international success.

In areas where we hand over control to AI systems, a risk analysis must be conducted, which is the core of the new EU AI Act that came into effect on 1 August this year. Developers of new AI systems must be able to document the associated risks.

Some AI systems are already being banned, such as linking facial recognition to public transport, which has been implemented in certain cases in China. 

The Nordic countries, with their long history of protecting personal privacy and ensuring good working conditions, have a special role to play in influencing international standards. 

To achieve this, it is necessary not only to develop national AI strategies but also a unified Nordic strategy. We gain greater influence by speaking with a collective voice than if we act individually.