Depends entirely on what people are doing with the model! If they're working in safety-critical areas and liberally applying AI without sufficient domain knowledge or oversight, then yes, absolutely, mistakes can be life-threatening.
Off the top of my head: healthcare (diagnosis, medication, treatment decisions), engineering (structural design, electrical systems), aviation, autonomous vehicles, industrial automation, chemical processing, power grids, cybersecurity of critical infrastructure, or even software controlling medical devices.
IMO you don't have to be physically operating heavy machinery to make a mistake that get someone killed. I would posit it has already happened and will continue to happen.
Off the top of my head: healthcare (diagnosis, medication, treatment decisions), engineering (structural design, electrical systems), aviation, autonomous vehicles, industrial automation, chemical processing, power grids, cybersecurity of critical infrastructure, or even software controlling medical devices.
IMO you don't have to be physically operating heavy machinery to make a mistake that get someone killed. I would posit it has already happened and will continue to happen.