Liability, Warranties and Insurance
Bringing a humanoid robot into your life comes with inherent risk. There is risk to your belongings, risk to your home and risk to yourself and potentially others. Although there are operational standards set forth in order to minimize the risks to others and myself, I decided to embark on a journey to insure myself from further liability, my home from damage and the robot itself.
When I filled out the lead form on the Insurance group’s website “I am getting a humanoid robot for my home and I’d like to discuss insurance,” my phone rang almost instantaneously. I didn’t actually think it was possible that it was the agency so I didn’t answer. I went to my University faculty meeting shortly thereafter and my phone rang again. I didn’t get a chance to answer. I went to have a tea with a friend at a cafe and my phone rang again. I didn’t answer.
I googled the number and saw that it led back to an agent at the insurance company who had already emailed me. They were asking for my date of birth. Perhaps they thought this was a young person pranking them with a wild request. I confirmed my date of birth and they paused, then invited me on a video call. I agreed and got on the video call to see a very curious face peering back at me through the camera.
“What do you mean, you’re getting a robot?”
“I mean I’m getting a robot.”
“Oh, I see.”
As the actual insuring of the robot would happen with a different, more experimental department’s input, we went on to discuss personal liability insurance, home insurance which covers electronics and legal insurance. I wound up with personal liability insurance, home insurance, a referral to the more experimental department and a lot to think about.
As for the actual robot itself, the initial discussion pointed towards there not being much value in insuring it on its own based on the fact that it has a year-long warranty. The conclusion was that anything the insurance company could cover would be covered anyway with the warranty by Unitree themselves. In this case, the coverage would be based on theft of the robot which would actually be covered in my home insurance under the burglary clause. The logistics of this make it nearly impossible and are not my main worry.
My main interest is based on my plans to bring forth some of the theoretical questions about robotic and AI liability into reality. If the robot breaks a glass in a restaurant (a very innocuous experiment), the idea of liability is not quite clear. This question would trigger an array of other questions including:
What mode was the robot operating in at the time?
Remote
Teleoperated
Autonomous
Who was controlling the robot if it was in Remote or Teleoperated mode?
Has the Robot been in the environment before?
What did it learn then?
Was the Robot operating using skills it came knowing or skills it acquired with me?
Was it reasonable to assume that a person wouldn’t have also knocked over the glass given where it was placed?
Who placed it there?
Did the breaking glass hurt anyone?
How was the incident handled?
Each different answer to these questions could send the process of legal or ethical liability down a different pathway worth exploring. The most interesting of which, to me, would be about the mode of operation. In gaining personal liability insurance, I would argue that the robot’s actions are my actions when it is being operated in remote or teleoperated modes. That in the same way that if I were to be wielding a remote controlled car and it caused a driver of a real car to swerve, I would be responsible for its actions. However, if the robot is in autonomous mode, I am not controlling its actions directly. If it comes from the factory and robotics team in China with certain movements already trained, would they be responsible if those movements resulted in property damage? Is the robot itself liable?
One could argue in any number of directions in this deliciously murky topic. But now that I am in possession of personal liability insurance and soon to be in possession of a humanoid robot, we will put these theoretical quandaries to task in the real world and begin to set some precedent and inspire further practical regulation!