MPs are using AI to help them write parliamentary motions - without disclosing it

An increasing number of parliamentary texts in the UK parliament are being written with assistance from Large Language Models (LLMs), without any disclosure that AI was used, according to researchers at Edinburgh Napier University and Chalmers University of Technology in Sweden.  

The international research team has developed and used AI detectors trained on thousands of political motions in the UK and Sweden. What surprised the researchers was not that politicians were using AI – but that not a single one disclosed this openly.

By analysing specific passages in around 4,200 of the political proposals submitted to the Westminster parliament during the period of the study, the researchers found that 243 contained AI-assisted content, with the percentage increasing year on year. Analysis of the Swedish Parliament, where voters go to the polls on 13 September, suggested that nearly 10 per cent of motions there had been produced with the aid of AI in the 2025-2026 session.

“Other research shows that the use of AI may change the intended ‘tone’ of a document, or even potentially introduce outright errors,” said Dr Peter Barclay from Edinburgh Napier University’s Computer Science group. 

Researchers have noted that AI-generated texts often seem persuasive and well written, meaning that such issues might be overlooked even by the original author. Moreover, undisclosed AI use may erode public trust in government, given that an Opinions and Lifestyle Survey by the UK Office of National Statistics in June 2026 showed that only 4 per cent of respondents would trust AI to carry out decision-making in government.

Calls for greater transparency

As the use of AI has become widespread in society, it has become common practice to label texts or images produced with the aid of AI, including in the media. The Artificial Intelligence Playbook for the UK Government recommends that officials should be open with the public about where and how algorithms and AI systems are being used in official duties. 

Andrea McGlinchey, associate researcher at Edinburgh Napier University, said: “While our research does not address regulatory frameworks and focuses only on evidence of undisclosed AI usage in the language patterns of parliamentary texts, our results suggest that greater transparency is required.” 

The results of the detectors are interpretable

Researchers trained one detector for English and one for Swedish, starting with around 1,400 parliamentary motions from each country written between 2014 and 2020, before generative AI tools were available. To generate equivalent texts using AI, they first had a chatbot summarise each original text and, based on that summary, create a new text of its own. Using large datasets containing labelled texts that were known to be either human-written or AI-generated, the researchers then created highly accurate classifiers able to determine where AI had been used somewhere in the writing process.

The study used an interpretable classifier developed at Chalmers University of Technology, meaning that the results can be explained by examining which words and combinations of words were used to identify AI-assisted texts – but this identification depends on overall statistical patterns, not just the presence of one or two telltale phrases.  

To check the accuracy of the classifiers, they were tested on a large number of parliamentary motions from 2021, before the advent of generative AI, with only one full-text false positive.

When motions from 2022 up until this spring were analysed, the researchers noted a clear increase in texts where AI had been used in some form. Having been around 3.5 per cent in 2022, this rose steadily to around 15 per cent of the Westminster motions studied containing some AI-assisted text during the 2026 parliamentary year.