Researchers discovered that politicians in the UK and Sweden are using artificial intelligence to draft parliamentary speeches and bills without informing voters about this practice. The study detected AI-generated text in up to 15% of the speeches analyzed in these two countries.
The research did not identify the specific parties or politicians who used this technology. However, researchers warn that the lack of transparency raises democratic concerns, since citizens do not know when the texts they hear or read were written by machines instead of humans.
The study also analyzed other countries, but the most significant results were observed in the UK and Sweden, where the frequency of text with AI markers was highest. Researchers highlight the need to establish clear guidelines on the disclosure of AI use in the legislative process.
This discovery comes in a context of growing adoption of generative AI tools across various sectors, including politics. Experts warn that the absence of rules on the identification of AI-created content may affect citizens' trust in democratic institutions.
