Norwegian prime minister announces policy at Oslo press conference, warning AI risks letting children 'skip important steps' in learning

Norwegian Prime Minister Jonas Gahr Støre speaks at a press conference in Oslo on Friday. [EPA]
Norwegian Prime Minister Jonas Gahr Støre speaks at a press conference in Oslo on Friday. [EPA]

Norway has announced a policy that in effect bans elementary school students from using generative AI tools.

Prime Minister Jonas Gahr Støre said at a press conference in Oslo on Friday that the measure would take effect from the new school year beginning in August, aimed at preventing AI from harming children's learning.

"Using AI increases the risk that children will skip important steps they must go through in their education," Støre said. "The most important thing at school is that our children learn to read, write and do arithmetic."

The government said AI use would be banned in principle for children in grades 1 through 7, covering ages 6 to 13, while middle school students aged 14 to 16 would be permitted to use AI cautiously under teacher supervision.

High school students aged 17 to 19 should learn how to use AI appropriately to prepare for future education and employment, the government added.

Norway has been integrating computers into classrooms since the 1990s and expanded tablet use after 2010, gradually reducing its reliance on printed books and handwriting.

The government said it also plans to introduce legislation to fund a reversal of that trend, expanding the use of printed textbooks in classrooms.

Norway banned smartphones in schools in 2024 following a broad decline in academic performance, and also expanded teachers' authority to enforce classroom discipline.

In addition, the government announced in April a plan to ban social media use for children under 16 to reduce young people's screen time.

Meanwhile, the OECD's "Digital Education Outlook 2026" said general-purpose generative AI can significantly boost performance outcomes but may hinder actual learning when used without educational design.

The report found that students who practiced math problems using general-purpose generative AI improved their performance by 48 percent during the practice phase. When tested without access to AI, however, those students scored 17 percent lower than peers who had studied using conventional methods. The OECD said "doing well does not mean learning has occurred," noting that delegating problem-solving to AI can impede concept formation and the transfer of learning.

The OECD described the phenomenon as "metacognitive laziness." The report said that when students offload cognitive tasks to general-purpose chatbots, their ability to monitor and regulate their own thinking weakens, engagement with learning declines and skill acquisition can be hampered over the long term.


killpass@heraldcorp.com