Florida launches prison probe into ChatGPT over FSU taking pictures : NPR

[ad_1] <img src="https://npr.brightspotcdn.com/dims3/default/strip/false/crop/5000x3333+0+0/resize/1100/quality/50/format/jpeg/?url=http%3A%2F%2Fnpr-brightspot.s3.amazonaws.com%2F2b%2F92%2Fb8d60d554b669b9070a5a7fb31e9%2Fgettyimages-2210117281.jpg" data-template="https://npr.brightspotcdn.com/dims3/default/strip/false/crop/5000x3333+0+0/resize/{width}/quality/{quality}/format/{format}/?url=http%3A%2F%2Fnpr-brightspot.s3.amazonaws.com%2F2b%2F92%2Fb8d60d554b669b9070a5a7fb31e9%2Fgettyimages-2210117281.jpg" class="img" alt="Police investigate the scene of a shooting near the student union at Florida State University on April 17, 2025 in Tallahassee, Florida. Two people were killed and five injured in the attack. Florida's attorney general is now investigating OpenAI because the alleged shooter used Chat" fetchpriority="high"/> Police examine the scene of a taking pictures close to the coed union at Florida State College on April 17, 2025 in Tallahassee, Florida. Two individuals have been killed and 5 injured within the assault. Florida's lawyer common is now investigating OpenAI as a result of the alleged shooter used ChatGPT to assist plan the assault. Miguel J. Rodriguez Carrillo/Getty Photographs disguise caption

toggle caption

Miguel J. Rodriguez Carrillo/Getty Photographs Florida's lawyer common is launching a prison investigation into ChatGPT and its father or mother firm OpenAI over claims that the accused gunman in a taking pictures at Florida State College final 12 months consulted the AI chatbot earlier than killing two individuals and injuring 5 extra.

The Republican lawyer common, James Uthmeier, mentioned at a press convention in Tampa on Tuesday that accused gunman Phoenix Ikner consulted ChatGPT for recommendation earlier than the taking pictures, together with what kind of gun to make use of, what ammunition went with it, and what time to go to campus to come across extra individuals, in line with an preliminary evaluate of Ikner's chat logs.

"My prosecutors have checked out this and so they've advised me, if it was an individual on the opposite finish of that display screen, we might be charging them with homicide," Uthmeier mentioned. "We can't have AI bots which might be advising individuals on easy methods to kill others."

OpenAI spokesperson Kate Waters mentioned in a written assertion to NPR: "Final 12 months's mass taking pictures at Florida State College was a tragedy, however ChatGPT will not be answerable for this horrible crime." She mentioned the corporate reached out to share details about the alleged shooter's account with legislation enforcement after the taking pictures and continues to cooperate with authorities.

Uthmeier's workplace is issuing subpoenas to OpenAI looking for details about its insurance policies and inside coaching supplies associated to person threats of hurt and the way it cooperates with and reviews crimes to legislation enforcement, relationship again to March 2024. On the press convention, Uthmeier acknowledged the investigation is coming into into uncharted territory and is unsure about whether or not OpenAI has prison legal responsibility.

"We're going to have a look at who knew what, designed what, or ought to have completed what," he mentioned. "And whether it is clear that people knew that the sort of harmful conduct may happen, that a majority of these unlucky, tragic occasions may happen, and however nonetheless turned to revenue, nonetheless allowed this enterprise to function, then individuals should be held accountable."

OpenAI's Waters mentioned that the chatbot "offered factual responses to questions with info that might be discovered broadly throughout public sources on the web, and it didn't encourage or promote unlawful or dangerous exercise."

She continued: "ChatGPT is a general-purpose instrument utilized by tons of of thousands and thousands of individuals day by day for professional functions. We work constantly to strengthen our safeguards to detect dangerous intent, restrict misuse, and reply appropriately when security dangers come up."

Ikner, 21, is dealing with a number of fees of homicide and tried homicide for the April 2025 taking pictures close to the coed union on FSU's Tallahassee campus, the place he was a scholar on the time. His trial is ready to start on Oct. 19. Based on court docket filings, greater than 200 AI messages have been entered into proof within the case.

Rising issues about AI chatbots The Florida investigation comes amid rising issues over the position of AI chatbots in mass violence. Uthmeier had already introduced a civil investigation into ChatGPT's position within the FSU taking pictures, which is ongoing, and attorneys for the household of one of many victims say they plan to sue OpenAI.

OpenAI is already dealing with a lawsuit from the household of a sufferer critically wounded in an assault in British Columbia in February 2026 that killed eight individuals and injured dozens extra. The alleged shooter mentioned gun violence eventualities with ChatGPT and was even banned from the platform months earlier than the taking pictures, however was in a position to evade detection and create one other account, OpenAI advised Canadian authorities.

The Wall Road Journal reported that OpenAI's inside programs flagged the account's posts and staffers have been alarmed sufficient to think about alerting legislation enforcement, however that the corporate determined to not. OpenAI has mentioned it's making modifications to "strengthen" its protocol for referring accounts to legislation enforcement within the aftermath of the Canadian taking pictures.

Lawsuits are additionally mounting towards OpenAI and different makers of AI chatbots alleging they've contributed to psychological well being crises and suicides. (OpenAI has mentioned the circumstances are "an extremely heartbreaking state of affairs" and that it is working with psychological well being consultants to enhance how ChatGPT responds to indicators of psychological or emotional misery.)

A wrongful demise lawsuit filed towards Google in March over the suicide of a Florida man accuses the corporate's Gemini chatbot of pushing the person to "stage a mass casualty assault close to the Miami Worldwide Airport [and] commit violence towards harmless strangers," in line with court docket paperwork.

In response to that lawsuit, Google mentioned: "Gemini is designed to not encourage real-world violence or counsel self-harm. Our fashions typically carry out properly in a majority of these difficult conversations and we commit vital sources to this, however sadly they are not good." The corporate added that on this particular case, Gemini had "referred the person to a disaster hotline many instances."