The Episode and the Source
According to a report in The Korea Herald, South Korean President Lee Jae Myung responded on social media to critics of a recent practice: the presence of top investigators and prosecutors at his cabinet meetings. In support of his position, he cited an exchange with Gemini, Google's generative artificial intelligence. In the post, the chatbot reportedly described the objections as sophistry that openly denies the president's constitutional authority to oversee the executive, warning that excluding investigators from the room would make them unaccountable. The article's author, Lim Woong, a professor at Yonsei University in Seoul, observes that no head of state had ever cited a chatbot to defend a government decision.
Why the Argument Does Not Hold Up
The central point raised by the article is not technical but procedural: anyone who uses these tools knows how much a language model's response depends on how the question is framed. By changing the premise, the model, or the timing of the request, one can obtain the opposite conclusion. Regular users know this and rarely treat a response as definitive truth. The country already has ministries that interpret legal rules and a large body of constitutional law scholars; according to the author, the way the president presented the exchange suggests that neither had been consulted beforehand, although he doubts that the decision was really made that way.
The president later explained that he had published the exchange to show how fragile the opposing arguments were, on both logical and factual grounds. For the author, however, when constitutional interpretation and political prudence intertwine, brandishing a chatbot's response to close a debate is a misstep: it makes the work of government seem trivial and invites ridicule.
The Political Reaction and the Issue on the Merits
According to The Korea Herald, the opposition issued a challenge along the lines of 'I asked AI too,' and the president's method became the laughingstock of the debate rather than the policy under discussion. The article adds that calling the criticisms sophistry did not help, because a serious argument deserves a serious response, based on interpretation of the rules, verified facts, and a fair rendering of the opposing position.
The question on the merits remains distinct: whether the heads of investigative offices can attend cabinet meetings is one thing; whether they should is another. Their presence alongside the president would inevitably cast doubt on the impartiality of the cases pursued by their offices. The author recalls that something can be legal and still unwise, and that with approval ratings falling, the episode appears to many voters as impatient sarcasm rather than considered argument.
The Lesson Beyond the Case
The article does not assign blame to AI: it considers it useful for summarizing large quantities of documents and identifying clauses in seconds, but recalls that a model reflects the assumptions contained in the prompt and tends to agree with the person writing, unless explicitly pushed to object. A plausible answer is not a justified answer and cannot confer legitimacy on a political decision or assume responsibility for it. In a democracy, the author writes, legitimacy should come from reasons that the public can examine and from officials who can be removed by vote.
