Artificial intelligence is already changing the way opinion pages are filled, and the awkward question now facing editors is not whether contributors use it, but how much they are allowed to rely on it. At Aftenbladet, Heidi Hjorteland Wigestrand argues that AI can turn a tedious piece into something even more tedious, just at greater speed, and that the real challenge for newsroom gatekeepers is deciding where assistance ends and substitution begins.

Her column reflects a wider uncertainty across Nordic media about what counts as acceptable help. Aftenbladet says it has started using detection tools to get a sense of how widespread AI use is in submitted opinion writing, while also warning that such systems can misread human work. That caution matters, because, as the newspaper notes, one submission from SMB Norge was flagged despite the organisation insisting the text was written by a person.

The debate is not confined to one newsroom. According to Aftenposten, Manifest Media claimed that Pangram identified fully AI-generated language in NRK material, including a memorial piece about businessman Stein Erik Hagen, while also raising questions about other texts linked to former NRK commentator Cecilie Langum Becker. The reporting did not establish those claims as settled fact, but it added fuel to a broader argument about disclosure, editorial standards and the limits of automated detection.

Journalisten, meanwhile, reported that Adresseavisen had to correct ten opinion texts after concerns about extensive AI involvement, with Pangram later assessing nine of the corrected pieces as entirely AI-written and one as almost entirely so. The paper said its internal review found the author had drafted the arguments, but AI had significantly reshaped the final wording. That distinction matters, because it captures the grey zone many editors now have to navigate: an idea may be human, while the prose that reaches readers may be largely machine-made.

Altinget has taken a harder line. Its debate editors have said they became uneasy with the volume of AI-generated submissions across their Nordic editions, arguing that such texts often sound polished but detached, and can weaken the human exchange that opinion pages are meant to foster. In response, the publication introduced clearer rules requiring disclosure of AI use and allowing editors to reject submissions that are almost entirely machine-written, or to ask for a human rewrite instead.

Aftenbladet appears to be moving in the same direction. Wigestrand suggests that the newspaper does not want to ban AI outright, but does want clearer boundaries before the technology becomes so good that it can produce passable readers’ letters on demand. Aftenposten’s newsroom guidance offers one model, stressing that editors remain responsible for everything published and that all final editorial judgements must be made by people. For now, that may be the only workable answer: AI can help shape a submission, but it should not be allowed to erase the voice, judgement and responsibility that make opinion writing worth publishing in the first place.

Source Reference Map

Inspired by headline at: [1]

Sources by paragraph:

  • Paragraph 1: [2]
  • Paragraph 2: [2]
  • Paragraph 3: [3]
  • Paragraph 4: [4]
  • Paragraph 5: [5]
  • Paragraph 6: [2], [7]

Source: Noah Wire Services