Following the tragic shooting at Florida State University last year, where Phoenix Ikner killed two people and wounded six, investigators made a startling discovery: the perpetrator had consulted ChatGPT to plan the attack. Before committing the crime, Ikner allegedly asked the AI for advice on weaponry, ammunition, and strategies to maximize casualties. This revelation has prompted Florida Attorney General James Uthmeier to launch a criminal investigation into OpenAI, raising the provocative legal question of whether a software developer can be held criminally liable for the actions of their AI model. While Uthmeier suggests that if a person had provided this information they would be charged with homicide, experts note that applying criminal law to automated software is uncharted territory.

The challenge in holding OpenAI accountable lies in the complexity of current law, which typically requires human intent—a factor absent in standard AI responses. While corporate criminal prosecutions have occurred in the past, such as the cases against Purdue Pharma or Volkswagen, those instances involved intentional decisions made by human executives. Legal scholars suggest that while negligence or recklessness could be potential paths for prosecution, the burden of proof required for a criminal conviction remains exceptionally high. As a result, many legal experts believe that while civil litigation or new federal regulatory frameworks may be more practical avenues for oversight, a criminal case against OpenAI would face immense procedural hurdles.