Politicians in Sweden and the UK appear to be increasingly using artificial intelligence (AI) to help write parliamentary texts without disclosing it, according to a new study.
Researchers from Chalmers University of Technology in Sweden and Edinburgh Napier University in the UK analysed 13,565 Swedish parliamentary motions and 4,209 UK parliamentary statements published between 2021 and April 2026 using AI detectors trained to distinguish between human and large language model-generated political writing.
The researchers found that about 15% of UK parliamentary statements analysed in 2026 contained at least one paragraph classified as AI-generated.
In Sweden, 9.4% of motions submitted during the 2025–2026 parliamentary year contained at least one paragraph classified as AI-assisted, equivalent to around 300 motions.
Yet researchers said they found no disclosure of AI use in any of the texts they identified.
That marks a sharp increase in both countries. The figure stood at about 3.5% in the UK in 2022 and 1.7% in Sweden in the 2022–2023 period.
Jan Riise, an MP for Sweden’s Miljöpartiet, the Green Party, and its spokesperson on AI and digitalisation, said he was “not surprised” by the findings in the Swedish case.
He told Euronews Next that the party uses AI for “research purposes” and has an internal policy urging caution in its use.
“I personally prefer to rely on my own words,” Riise said.
He said politicians should label the use of AI.
“If you need to write an entire proposal, that’s another story,” he said, adding that he had “never asked [AI] to write the whole thing”.
Calls for transparency
When the researchers assessed entire texts rather than individual passages, 6.5% of Swedish motions were classified as AI-generated during the latest parliamentary year, compared with 2.1% of British texts in 2026.
Researchers behind the study are calling for more transparency.
“Regardless of the extent to which AI has been used, there may be risks associated with a lack of transparency,” said Minerva Suvanto, the study’s lead author and a PhD student at Chalmers University of Technology, in a statement.
“We also note that texts produced with the help of AI are often very well-written and persuasive, which increases the risk that errors will not even be spotted by the writers themselves,” Suvanto added.
How researchers detected texts generated by AI
To detect possible AI use, researchers developed separate systems for Swedish and English parliamentary texts.
They trained them on human-written texts from before generative AI became widely available, alongside AI-generated versions on similar political topics.
The detectors were then applied to texts published from 2021 onwards, analysing both full documents and individual paragraphs. A text was counted as containing AI-assisted content if at least one sufficiently long paragraph was classified as AI-generated.
When tested on parliamentary material from 2021, before ChatGPT became publicly available, only one full text in each country was incorrectly classified as AI-generated, according to the study.
“We use what is known as interpretable AI,” said Mattias Wahde, professor of applied artificial intelligence at Chalmers University.
“This means that we can not only investigate whether AI has been used but also where it is evident in the text. But you can’t say that individual words are ‘AI words’ — it’s more complex than that.”
However, the study cannot establish which AI tool was used or exactly how it was used in drafting the texts.
The findings come as some politicians are acknowledging the use of generative AI in their work.
Swedish Prime Minister Ulf Kristersson revealed in August 2025 that he regularly used tools including ChatGPT and the French chatbot LeChat to seek a “second opinion” on political questions.
He said he used AI to ask what others had done and whether Sweden should consider doing the opposite. His press secretary later said no security-sensitive information was entered into the tools.
