OpenAI addressed security concerns after Las Vegas law enforcement revealed that ChatGPT was used to plan a January 1st incident at Trump International Hotel. This marked the first known case of the AI platform being used in such a context on U.S. soil.
Las Vegas Metropolitan Police Department officials confirmed on Tuesday that a suspect used ChatGPT, a popular AI chatbot, in attempting to determine explosive quantities for an incident involving a Tesla Cybertruck outside the Trump International Hotel on New Year's Day.
The FBI's investigation determined that the suspect, identified as Matthew Livelsberger, 37, an active-duty Army soldier from Colorado Springs, acted alone. Seven people sustained minor injuries in the incident, which authorities have characterized as a suicide. Investigators noted that Livelsberger likely suffered from post-traumatic stress disorder and had no apparent animosity toward President-elect Donald Trump.
This unprecedented use of ChatGPT in a security incident raises immediate questions about AI safety protocols and their effectiveness in preventing misuse. OpenAI's response highlights their existing safety measures while acknowledging the limitations of controlling publicly available information.