AI Criminal Liability Still Controversial, Experts: Establishing Regulatory System is Key
Amid ongoing debate over whether AI should bear criminal responsibility, legal experts argue that establishing a comprehensive regulatory system is more crucial than pursuing criminal charges, to address gaps in existing laws concerning emerging AI technologies. This discussion was sparked by a case where a suspect consulted an AI chatbot before a school shooting.
📋 Article Processing Timeline
- 📰 Published: May 11, 2026 at 13:37
- 🔍 Collected: May 11, 2026 at 14:01 (24 min after Published)
- 🤖 AI Analyzed: May 11, 2026 at 22:59 (8h 57m after Collected)
Central News Agency
(Central News Agency, New York, 10th, comprehensive foreign report) Amid the ongoing debate over whether artificial intelligence (AI) should bear criminal responsibility, legal experts point out that instead of pursuing accountability through criminal litigation, it is more important to establish a comprehensive regulatory system to fill the gaps in existing laws regarding emerging AI technologies.
According to an AFP report, before last year's Florida State University campus shooting, suspect Phoenix Ikner had conversed with an AI chatbot, not friends, parents, or anyone who might have stopped him.
According to evidence obtained by Florida prosecutors, the student had asked ChatGPT which weapons and ammunition were best for an attack, and when and where to cause the most casualties. Investigators said the chatbot answered his questions.
Now, State Attorney James Uthmeier wants to clarify whether this means OpenAI bears criminal responsibility.
In announcing a criminal investigation into ChatGPT developer OpenAI, he stated, "If there was a person on the other side of the screen, we would charge them with murder," and he did not rule out the possibility of criminal charges against the company or its employees.
This campus shooting, which occurred in April 2025, brought a highly controversial issue into legal focus: whether AI developers should bear criminal responsibility for AI's role in crimes or suicides. Experts point out that this is legally feasible but an extremely complex issue.
●Criminal Product?
Under U.S. law, it is indeed possible for relevant companies to face criminal prosecution, but such cases are still rare.
Late last month, Purdue Pharma was fined over US$5 billion in criminal penalties for fueling the opioid crisis; Volkswagen was found guilty in the emissions cheating scandal; and Pfizer and Exxon have also been held accountable for various incidents.
However, the above cases all involved human decision-making, while the core of this case lies in the AI's own actions, which is where the law is most tricky.
Matthew Tokson, a law professor at the University of Utah, said: "Ultimately, it's the product that facilitated the crime, which makes this case both unique and quite tricky."
Legal experts point out that companies are more likely to face charges of "negligence" or "recklessness," which are usually misdemeanors with lighter penalties.
OpenAI believes that ChatGPT should not be held responsible for this and emphasizes that the company will continue to strengthen safety mechanisms to detect and prevent abuse.
●Civil or Criminal?
For those seeking accountability, civil litigation may be a more viable path.
Tokson said that civil litigation could force companies to be more cautious in developing products and to face the human consequences of their mistakes.
Currently, there are multiple civil lawsuits against AI platforms in the U.S., many of which involve suicide cases, but so far, there have been no rulings holding relevant companies responsible.
Matthew Bergman, an attorney at the Social Media Victims Law Center, pointed out that the new version of ChatGPT has added more safety mechanisms, but it is still "not enough to fully protect."
However, Brandon Garrett, a law professor at Duke University, believes that instead of relying on criminal prosecution, it is "wiser" to establish a comprehensive regulatory system through Congress and the government. (Compiler: Liu Wen-yu) 1150511
Choose to stand with facts, your every sponsorship is the power to protect press freedom.
Download the Central News Agency's "First-hand News" APP to stay updated with the latest news instantly.
The text, images, and audio/video on this website may not be reproduced, publicly broadcast, publicly transmitted, or used without authorization.
(Central News Agency, New York, 10th, comprehensive foreign report) Amid the ongoing debate over whether artificial intelligence (AI) should bear criminal responsibility, legal experts point out that instead of pursuing accountability through criminal litigation, it is more important to establish a comprehensive regulatory system to fill the gaps in existing laws regarding emerging AI technologies.
According to an AFP report, before last year's Florida State University campus shooting, suspect Phoenix Ikner had conversed with an AI chatbot, not friends, parents, or anyone who might have stopped him.
According to evidence obtained by Florida prosecutors, the student had asked ChatGPT which weapons and ammunition were best for an attack, and when and where to cause the most casualties. Investigators said the chatbot answered his questions.
Now, State Attorney James Uthmeier wants to clarify whether this means OpenAI bears criminal responsibility.
In announcing a criminal investigation into ChatGPT developer OpenAI, he stated, "If there was a person on the other side of the screen, we would charge them with murder," and he did not rule out the possibility of criminal charges against the company or its employees.
This campus shooting, which occurred in April 2025, brought a highly controversial issue into legal focus: whether AI developers should bear criminal responsibility for AI's role in crimes or suicides. Experts point out that this is legally feasible but an extremely complex issue.
●Criminal Product?
Under U.S. law, it is indeed possible for relevant companies to face criminal prosecution, but such cases are still rare.
Late last month, Purdue Pharma was fined over US$5 billion in criminal penalties for fueling the opioid crisis; Volkswagen was found guilty in the emissions cheating scandal; and Pfizer and Exxon have also been held accountable for various incidents.
However, the above cases all involved human decision-making, while the core of this case lies in the AI's own actions, which is where the law is most tricky.
Matthew Tokson, a law professor at the University of Utah, said: "Ultimately, it's the product that facilitated the crime, which makes this case both unique and quite tricky."
Legal experts point out that companies are more likely to face charges of "negligence" or "recklessness," which are usually misdemeanors with lighter penalties.
OpenAI believes that ChatGPT should not be held responsible for this and emphasizes that the company will continue to strengthen safety mechanisms to detect and prevent abuse.
●Civil or Criminal?
For those seeking accountability, civil litigation may be a more viable path.
Tokson said that civil litigation could force companies to be more cautious in developing products and to face the human consequences of their mistakes.
Currently, there are multiple civil lawsuits against AI platforms in the U.S., many of which involve suicide cases, but so far, there have been no rulings holding relevant companies responsible.
Matthew Bergman, an attorney at the Social Media Victims Law Center, pointed out that the new version of ChatGPT has added more safety mechanisms, but it is still "not enough to fully protect."
However, Brandon Garrett, a law professor at Duke University, believes that instead of relying on criminal prosecution, it is "wiser" to establish a comprehensive regulatory system through Congress and the government. (Compiler: Liu Wen-yu) 1150511
Choose to stand with facts, your every sponsorship is the power to protect press freedom.
Download the Central News Agency's "First-hand News" APP to stay updated with the latest news instantly.
The text, images, and audio/video on this website may not be reproduced, publicly broadcast, publicly transmitted, or used without authorization.