Back in 1949, a comic strip named Ella Cinders (Cinderella, geddit?) had its plucky heroine find the missing half of a lost blueprint from the late practical joker Cromwell Shnook. When the device was built it turned out to be a robot, but nobody knew what it was supposed to do. All Cromwell’s will says is “I want my nephew Horace to press the top button. I am sure he’ll get the surprise of his life.”
He sure does. The robot kicks him in the seat of the pants, something Uncle Cromwell always wanted to do.
Assume this happens in real life. Does Horace have any recourse for this? Who, if anybody, has the liability for the robot’s actions? What if the robot strangled Horace? He pressed the button that started the action. What if the robot were then to strangle someone else after Horace’s death?
My untrained notion is that the company that blindly followed the blueprints would in today’s world get sued into oblivion but the legal trail of liability isn’t as clear.
Legal liability generally flows at least partly from the idea of foreseeability. In something like a Rube Goldberg machine, the whole point of the joke is that the outcome is foreseeable, even if it takes a lot of work to figure out how the cat makes the bowling ball open the canary’s cage using a ceiling fan. A Rube Goldberg device with a big inscrutable black box in the middle wouldn’t be entertaining.
Your case essentially demands that nobody charged with building the device from the plans has any insight into what it will do when switched on. At least so far in history our machines don’t work that way. The plans make it pretty obvious to anyone skilled enough to build the thing what it’ll probably do or not do.
Admittedly at some point in the future with AI or at least A-quasi-I it will become increasingly true that only generalities about behavior can be predicted, and not even that with absolute certainty.
Here’s another take completely:
At the risk of launching a hijack from the very first response, consider the law of firearms and liability. It’s fully foreseeable that a firearm will injure or kill someone if fired at them. Legal technicalities aside, it’s also pretty obvious from the plans whether or not any given gizmo is a firearm or not.
Yet the general common-law legal perspective is the manufacturer has no liability if the thing is used as designed to kill something and that something happens to be human.
On a less controversial note a car is clearly a dangerous thing. A person can deliberately or inadvertently do a lot of harm with one. Despite, unlike firearms, that harmful outcome being completely outside the car’s design goals. And again the law says the user bears all responsibility for any damage their use of the (non-defective) tool causes, while the manufacturer bears none.
Back at the OP, that would imply Horace pushed the button and Horace owns the consequences. Maybe he shoulda thought twice before pushing it. Or at least he should’ve tried to figure out what will happen next.
I’m assuming the button is on some part of the robot or its paraphernalia. That’s a leap, but IMO a real small one. If the button’s not there, where is it?