Whose fault is greater: the boyfriend who took the first bite of the chicken drumstick, or the girlfriend who got upset and angry about it? Between the partner who canceled a date saying it was too hard and the partner who demanded keeping the plan without considering the circumstances, who bears more responsibility?
More people are now asking artificial intelligence (AI) about issues they once would have confided to friends or family or resolved between partners. AI services have even emerged that analyze points of contention and responsibility in relationship conflicts and present opinions in the form of a written judgment.
According to social media (SNS) on the 6th, the app "Nae Marimajji," which resolves conflicts between lovers in a trial format, has recently been drawing attention. Named to sound like "I'm right," the app lets one side write up the conflict, the other side post a response, and the AI organizes both sides' claims and facts to present an opinion in judgment form.
◇Even trivial dating dilemmas go to the "AI court"
Stories posted to the app are things that commonly happen between lovers. From taking the first bite of a chicken drumstick to sitting on the bed in outside clothes or putting off the dishes for two days, all have become subjects for the AI court's ruling.
One user said the boyfriend secretly checked the user's phone while the user was asleep, then blocked changing the password and inspected the phone whenever the user stepped away. The AI said, "The other party must respect the applicant's privacy and immediately stop behavior that causes anxiety," adding, "They should sincerely apologize and make efforts to rebuild mutual trust."
There was also a conflict over canceling plans. The boyfriend said, "Even if plans are set, they can be canceled on days when the body and mind are exhausted." The girlfriend countered, "Unless you're truly sick, you should keep your promises."
The AI framed the issue as whether to prioritize the obligation to keep promises or the need to consider the other person's condition. It concluded, "There is greater responsibility in the attitude of forcing the fulfillment of a plan without fully understanding the other person's fatigue and need for rest."
However, it also said the boyfriend needs to clearly explain his condition and needs, and try to understand the girlfriend's feelings of disappointment over the cancellation. Rather than naming a winner, the conclusion was to acknowledge differences in perspective and adjust how they communicate.
The developer of Nae Marimajji said the AI is not an absolute judge that hands down the right answer on relationships, but a communication tool that helps lovers, whose emotions are running high, organize their positions in writing and restart a conversation.
◇Dating advice, restaurant picks, workout plans… AI in everyday life
Beyond dating advice, the use of Generative AI is rapidly expanding across daily life, from restaurant recommendations and workout guidance to document drafting and translation.
The Ministry of Data and Statistics (MODS) plans to newly include "software subscription fees," which cover Generative AI services, in the representative items of the consumer price index in December. It means Generative AI has moved beyond a novelty used by a few to a life service for which the average household continually incurs expense and expenditure.
According to the "2025 internet usage survey" by the Ministry of Science and ICT, Korea's experience rate with Generative AI is 44.5%, up 11.2 percentage points from a year earlier. Paid subscription reached 7.9%.
A graduate student surnamed Kim (28), who said the person often uses Gemini, said, "If I ask AI about a dilemma, I can get an immediate answer without exposing private matters to those around me," adding, "It also helps a lot with saving time when handling work."
◇The more we rely on AI, the less we "think"… "Guidelines needed"
However, as AI's scope widens, concerns are growing about excessive reliance. Recently in the legal community, there were cases where lawyers and parties to cases submitted fake precedents generated by AI to courts and investigative agencies without verification.
Park, a 38-year-old in charge of external cooperation at work, said, "A team member brought a proposal written with Generative AI, but the mechanical tone and factual errors were barely reviewed, so I had to rework it from scratch," adding, "Using AI itself is efficient, but using the output as-is without verification is another problem."
In fact, an experiment on 88 college students by researcher Ju So-mi and professors Lee Chang-jun and Lee Dae-ho of the Department of Artificial Intelligence Convergence at Sungkyunkwan University found that the higher the reliance on AI, the more the process of thinking independently and verifying content diminished. Students with high AI reliance had lower actual test scores but showed so-called "learning illusion," perceiving that "studying went well."
As AI's scope of use expands rapidly, some say society needs discussion and education on how to use it properly. Lee Chang-jun, a professor at Sungkyunkwan University, said, "We need educational guidelines so Generative AI is used properly as an assistive tool to support thinking, not as a tool that replaces judgment."