The moment Michael Butler pressed the accelerator—or didn't—a woman died inside her own home.
The real question isn't whether the driver failed or the automation failed—it's whether Tesla's name for this feature promised something the system cannot do. Whether that promise created the exact conditions for this death to happen.
FSD is not full self-driving. This is not controversial among engineers who work on autonomous systems—it requires active driver supervision, fails in ways that don't announce themselves. Becomes unreliable in the presence of obstacles, pedestrians, or scenarios outside its training data.
When a product's name describes a capability it doesn't possess and a death follows, the law has a mechanism for that called product liability. It doesn't care whether the driver was paying attention—it asks whether the product was honestly marketed, whether a reasonable consumer would understand what they were buying. Whether the gap between promise and reality created a foreseeable risk of harm. Butler believed—or at least claimed to believe—that his car could drive itself. That belief had to come from somewhere other than the technical specifications buried in the manual.
The criminal court is asking who failed to control the car. Product liability asks whether the car was honestly sold.
”Even if the logs show the car did everything correctly and Butler was simply not paying attention, the liability question remains untouched—a product that requires constant supervision cannot be branded as though it doesn't. The fact that most drivers using FSD are still alive is not evidence the system is safe. It's evidence they've been lucky.