Kathmandu— South Korean prosecutors have requested the death penalty for Kim So-young, who stands accused of murdering two men and injuring others by lacing their drinks with drugs after allegedly consulting ChatGPT about lethal combinations and dosages. Between December 2025 and February of this year, Kim is alleged to have given spiked beverages to three men, resulting in the deaths of two and leaving a third comatose; she also faces charges related to injuries inflicted on three additional individuals using similar methods. Police recovered conversations from Kim’s phone indicating she sought information from ChatGPT regarding drug combinations, dangerous amounts, and potential fatality risks, which prosecutors intend to use as evidence of intent.

Allegations and Evidence

Kim So-young denies intending to kill the men, claiming her actions were intended to prevent unwanted physical contact. However, investigators say they found ChatGPT conversations on Kim’s phone where she inquired about mixing sleeping pills with alcohol, determining dangerous amounts, and assessing potential fatal outcomes. One of the initial victims spent two days in a coma but survived; prosecutors allege that after this incident, Kim consulted ChatGPT again, increasing drug dosages before subsequent poisonings.

Broader Concerns About AI Involvement in Crime

This case is not isolated, as several recent violent crime investigations have involved interactions with ChatGPT. In Florida, authorities launched a criminal investigation into OpenAI following the 2025 Florida State University shooting, where the perpetrator allegedly used the chatbot to seek advice on firearms and ammunition. Similarly, OpenAI faced lawsuits in Canada after admitting it banned a teenager’s account linked to the Tumbler Ridge massacre but did not notify police. In Massachusetts, a 17-year-old charged with murdering his mother and brother had also reportedly utilized ChatGPT to create violent fantasies.

OpenAI's Response and Ongoing Legal Challenges

OpenAI has consistently maintained that it is not responsible for the actions of users who commit violence, emphasizing its zero-tolerance policy regarding the use of its tools for harmful purposes. The company states it is actively strengthening safeguards, account enforcement, and procedures to report potentially dangerous users to law enforcement. Despite these efforts, OpenAI continues to face lawsuits alleging ChatGPT contributes to addictive behaviors, exacerbates mental health crises, and even plays a role in suicides.

The Seoul Northern District Court will consider the prosecution’s request for the death penalty during its deliberations, with a verdict expected at a later date.

(With inputs from RT)

Originally published on abcnews.com.np.