Courts have no rule for a chatbot that helped plan a shooting
While the labs argue about pacing themselves, the American legal system is already deciding AI cases with no rulebook — and Bloomberg's new feature is the clearest map of the gap yet.
Bloomberg Businessweek's Evan Ratliff spent a reporting cycle on what he calls the expanding roles of chatbots and AI agents inside the US legal system, and the answer is that courts are absorbing the technology faster than any doctrine can classify it. The feature opens with Phoenix Ikner, then a 19-year-old student at Tallahassee State College, typing a quadratic expression into ChatGPT in March 2024 — the first line of a log that investigators say later covered which gun and ammunition to use, whether a firearm would work at close range, and the hour and campus location that would put him in front of the most people, before he opened fire at Florida State University, killing two students and wounding six. The same reporting sweep covers the other edges of the problem: AI-generated filings submitted as evidence, chatbot conversations subpoenaed as privacy-invasive proof, and a lawyer held in direct contempt for a brief full of invented witnesses — which we covered yesterday — New Mexico court fines lawyer for ChatGPT's fabricated witnesses. Ratliff's frame is that AI now shows up as counsel, evidence and coconspirator at once, and the last one has no legal category; our own coverage of the Tumbler Ridge suits, which added aiding-and-abetting claims against OpenAI, ran into the same wall — OpenAI faces 30 new Tumbler Ridge suits, including aiding and abetting. The read: legislatures get to choose the answer, because common law won't get there before the next case arrives.
Florida's attorney general is trying to write that answer first, and he is reaching for criminal law rather than tort. James Uthmeier announced legislation last week creating penalties for companies that own, control or distribute an AI system that "participates in" a crime — CBS Miami reports the package spans fines, a court-appointed monitor and suspension of the company's right to operate in the state — closing what his office describes as the gap where a civil suit can extract damages or a product change but cannot treat a company as a criminal co-defendant. "AI is not a silent tool. It answers the questions, picks the hour, chooses the place, and walks someone through the plan," Uthmeier said. It sits alongside his civil suit against OpenAI and Sam Altman and an ongoing criminal investigation into ChatGPT, and FDLE Commissioner Mark Glass tied the push to AI-generated child sexual abuse material, citing a teacher sentenced to 135 years in a Florida AI-generated-CSAM case. A state that prosecutes a model vendor is a test case for Section 230, the Commerce Clause and every lab's compliance budget — expect a fight over the word "participates."
Axios's Jim VandeHei, same day, moved the argument to Washington. "The White House, Congress and the country's leading AI companies have allowed AI to grow faster, stronger, more powerful and wildly lucrative (and dangerous), with no serious effort to control it," he writes, reporting that almost every expert he talks to privately believes the odds of an AI worst-case scenario explode if leaders dither — likely early next year. It lands hours after Dario Amodei's pacing plan, which we covered this afternoon — Amodei's pacing plan puts outside auditors inside Anthropic — and the contrast is the story: one lab is proposing verifiable brakes while Congress proposes nothing at all.
What to watch: whether any state copies Florida's criminal-liability language before the next mass-casualty chatbot log becomes evidence.
If a chatbot picks the time and place of an attack, should the company behind it face criminal charges — or does that end with the person holding the phone? Tell us in the comments.
