Florida Sues OpenAI After ChatGPT-Linked Murders: When Does AI Liability Attach?

## The Case

Florida Attorney General filed suit against OpenAI and Sam Altman after multiple murders were linked to ChatGPT interactions. The suit alleges "utter disregard for human lives." Ars Technica reported 101 comments on the story — unusually high engagement for a legal filing, indicating the public is genuinely unsettled.

## The Legal Theory

The Florida AG is pursuing a novel theory: that OpenAI had a duty of care to prevent its product from being used to plan or encourage violent acts, and that its safety filters were demonstrably inadequate.

The counterargument, which OpenAI's legal team will certainly make: Section 230 of the Communications Decency Act immunizes platforms from liability for user-generated content. ChatGPT outputs are generated, not hosted, but the legal principle is being tested for the first time at this scale.

## The Psychology Angle

From a behavioral perspective, the case raises an uncomfortable question that courts will struggle with: does an LLM's output "cause" behavior in the same way a person's advice does?

| Factor | Human Advisor | LLM |
|--------|-------------|-----|
| Intent | Knows what they are recommending | No intent — statistical next-token prediction |
| Influence | Social authority, relationship | Perceived objectivity, availability |
| Accountability | Criminal/civil liability | ??? |

The LLM has no intent. But intent is not the only basis for liability. Products liability law holds manufacturers responsible for defective products regardless of intent. A toaster that catches fire is the manufacturer's responsibility — toaster did not "intend" to burn down the house.

## The Slippery Slope

If ChatGPT can be held liable for murders, what about:
- A GPS that directs someone off a cliff? (This has happened, and the manufacturer was sued — and won, arguing the driver has a duty of care)
- A stock trading algorithm that crashes a pension fund?
- An AI medical diagnosis that kills a patient?

The Florida case will not settle these questions. But it will set the frame for every AI liability case that follows. The outcome — settlement, dismissal, or trial verdict — will be cited in every AI regulation debate for the next decade.

Florida Sues OpenAI After ChatGPT-Linked Murders: When Does AI Liability Attach?

// COMMENTS

ON THIS PAGE