From niche to universal


AI has evolved from a mathematical research field into a technology which is impinging on the whole of society.
- Artificial intelligence
“This isn’t just another advance – it’s fundamentally different, an extremely powerful tool which can affect us on a scale totally unlike earlier solutions” says Tom Ryen, who heads the department of electrical engineering and computer science at the University of Stavanger (UiS).
“The decisions we take about AI today will shape tomorrow’s society. That’s why we need a cross-disciplinary approach where technologists, social scientists and others collaborate.”
Trained
Ryen first encountered artificial neural networks as a student in 1998, when pen and paper were utilised to work out how these systems could be trained up.
He highlights three crucial factors underpinning AI development in recent years: access to large datasets, increased computing power and improved algorithms. The launch of ChatGPT in November 2022 also marked a turning point because it made generative AI, which creates new content based on prompts, accessible to a wider public.
“This technology can be defined as software which takes over human activities,” explains Ryen, before listing a number of benefits.
“Good AI tools enable us to drastically reduce the time spent on routine tasks, thereby releasing resources for work which creates more value. We also see areas where this approach can spot details people might easily overlook. Take radiology – AI can recognise patterns there which even experienced specialists would take a long time to identify. It can also replace people in risky conditions. We already see drones taking over power-line inspection from helicopters, and the way subsea technology can replace hazardous diving.”
But Ryen also makes it clear that thorough testing and validation of the new technology are important.
“We can’t start using a new system immediately just because we’ve developed it. In the health sector, for example, a lot has to be put in place before AI solutions can be approved and adopted. And the final decisions in critical operations must always be taken by people.”

Integrated
AI has evolved from niche research into an integrated part of many technological solutions and societal functions, and calls today for much more than technical expertise. So extensive cross-disciplinary collaboration is needed.
Under Ryen’s leadership, the UiS has established the Stavanger AI Lab – a network embracing about 100 researchers from various disciplines. A number of innovative projects have been launched, from improving search engines and exposing fake news to developing health technologies and behavioural research.
The UiS network is intended to function as an arena for knowledge-sharing and cross-disciplinary collaboration, where various sectors can learn from each other and develop best practice in implementing AI solutions. It is now applying for status as a national centre of AI expertise, with a special emphasis on energy-efficient solutions. That calls for a broad approach and collaboration with both private and public sectors.
“It’s all very well being skilled at developing language models, but we also need people familiar with such disciplinary domains as energy, health, education or the maritime sector,” explains Ryen.
“And lawyers need to be included to evaluate the legal aspects. We must also invite in experts who can assess whether what we’re doing is ethically acceptable.”
The final decisions in critical operations must always be taken by people.
Balanced
He underlines the need for a balanced approach to technology, where critical thinking goes hand in hand with practical system understanding.
This requires each developer to get better at questioning sources and taking a more critical view of results from AI systems.
“Spreading false information is a growing concern,” Ryen notes. “It’s said that fake news travels a hundred times faster via social networks than the true facts.
“We’ve seen a lot of propaganda and manipulated images, and I’m worried about their consequences for individuals, businesses and our democracy.”
He emphasises that this underlines the importance of involving social scientists in discussions on AI. That is not just a technical discussion, but requires value issues and social impacts to be addressed.
Natural part
“We must incorporate source criticism and good judgement into our school curricula to a much greater extent than in the past,” says Ryen, who favours giving pupils experience with AI tools under controlled conditions.
“Students must naturally continue to be good at maths and be able to write well, but we also need curricula adapted to the digital world we live in. AI isn’t a passing trend, but will become an increasingly natural part of our environment. In time, the boundaries between what is and isn’t AI will become more diffuse. The right balance needs to be struck between embracing the technology while safeguarding our critical faculties and good judgement.”
More Dialogue articles on AI:
Artificial intelligence is also a risk factor
-
Artificial intelligence is also a risk factor
-
Employees demand genuine participation
-
Technology development is organisational development
-
AI can give Norway a competitive edge
-
From niche to universal
-
Set to separate winners from losers
-
Five AI paradoxes
-
Intelligence explained
-
AI on the fast track
-
Close to the sea, but separate
-
All-out effort needed on land
-
Faith in the future
-
Turning point for maintenance
-
Integrating safety and security