Transparency Concerns Arise Over Undisclosed AI Use in Swedish Parliamentary Motions

A recent study conducted by Chalmers University of Technology and Edinburgh Napier University has revealed that nearly 10% of parliamentary motions in Sweden’s Riksdag were drafted with the assistance of AI chatbots over the past year. However, none of the politicians have disclosed their use of AI, prompting significant concerns about transparency in the political process.

Researcher Minerva Suvanto, a doctoral student at Chalmers, pointed out the critical transparency gap, stating, "There is a lack of transparency," which could become a "trust issue in the long run." The study tracked AI usage in motions over the last four years, finding a steep rise from almost no AI-assisted motions to about 300 in the 2023-2024 parliamentary year. Specifically, 1.9% of motions were entirely AI-assisted, and 4.1% included at least one AI-assisted section.

The researchers developed specialized AI detectors that found no significant variation in AI usage among different political parties or regions. However, it remains unclear whether AI was utilized for drafting entire texts or just for tasks like grammatical editing. Mattias Wahde, a professor at Chalmers, stressed that the key issue is not merely the growth in AI use but the complete absence of disclosure by politicians, which may damage voters’ trust in political transparency.

Additionally, studies from the SOM Institute in Gothenburg highlight that voters view transparency about AI use as highly relevant. The findings suggest the urgent need to incorporate clear disclosure of AI assistance in motions to uphold responsible governance and maintain public trust.