Swedish Prime Minister Ulf Kristersson has revealed that he frequently consults artificial intelligence tools such as ChatGPT and the French platform LeChat to test and evaluate his ideas before making decisions. In an interview with Dagens Industri, Kristersson explained that AI helps him explore alternative approaches and assess how similar issues have been addressed in other contexts. He stressed, however, that no sensitive government data is ever shared with these systems.
AI in Swedish political decision-making
Kristersson described AI as a useful tool for handling the large volumes of information involved in government work. According to him, both ministers and political advisers can benefit from AI to better analyse and verify information. Still, he pointed out a significant limitation: the need for language models trained more extensively on Swedish experiences and data.
Debate over security and political implications
The Prime Minister’s remarks have sparked a political debate in Sweden. The centre-left newspaper Aftonbladet criticised the practice, calling AI an “unprofitable” technology driven by big tech speculation. The editorial warned that relying on such tools could be akin to “lending one’s brain to a machine” and raised national security concerns, noting that queries submitted to AI systems may be stored on US servers and potentially linked to the user.
Expert concerns about AI bias and reliability
Virginia Dignum, professor of responsible artificial intelligence at Umeå University, told The Guardian that AI cannot provide genuine political opinions because it reflects the biases and training data of its developers. She warned that leaders risk placing undue trust in systems that simulate intelligence without true understanding.
