How AI Models Would Vote in Sweden

Nordan AI published a study examining which Swedish political parties large language models would choose, apparently using voting-style questions or political questionnaires to probe model preferences. The article was shared on Hacker News, where it had a score of 4 and one comment at the supplied metadata snapshot. No model list, prompts, sample size, aggregation method, or results are available here, so the claims should be treated as a research report requiring methodological verification rather than evidence that models possess stable political beliefs.