When it comes to employment law, I am a robot skeptic. That is, I think employment lawyers ought not worry too much about robotics.
My skepticism is rooted in the vagueness of the concept of a “robot.” A recent Harvard Business Review article, What is a Robot Anyway, tackled this definition problem. Are they industrial, service oriented, or both? Can they be virtual or only actual, or both? Are they necessarily mobile or not? Ultimately, the article’s effort fades…we get nothing satisfactory.
A diversion – kindly skip to the next paragraph if you didn’t like philosophy in college. There is a principle in philosophy called supervenience. Supervenience, according to that oracle of philosophical knowledge Wikipedia, is “used to describe cases where (roughly speaking) the lower-level properties of a system determine its higher level properties.” If the higher level properties are wholly dependent on the lower level ones, then it is questionable that the higher-level ones are meaningful. In other words, if a thing can only be meaningfully defined as a collection of its parts, then the status of the whole is questionable. If the whole is somehow greater than its parts, them it may be worth talking about.
In robotics, the lower level properties always define the higher level ones: a robot is just a collection of different pieces (computational power, software, actuators, interfaces, etc). When we, as lawyers, struggle to define what a robot is, we are struggling to understand which of those different pieces need to be present in order for us to have a robot.
I think it may be smarter from an employment law point of view to reject the idea of robots and instead focus on the parts. Does your computing platform move? Worry about workplace injuries. Does your autonomously moving system collect data? Worry about inappropriate data collection and eDiscovery. Does your HR expert system interact with potential employees? Worry about whether it is not asking proper questions. Displacing workers? Worry about union responses.
In short, the form factor – or even the question of whether the device is actual or virtual – is irrelevant. What matters are the parts: the mobility, the data collection, the human resources impact, the computing behind the “robot’s” decision-making.
A robot is just a collection of legally (very) significant parts. By focusing on the whole instead of the parts, we lose focus on what will matter in the workplace.
One caveat: I think we will come to a meaningful definition of robotics in human augmentation, most interesting for employment law purposes as devices such as mechanical exoskeletons and other assistive technology change how we view accommodations.
Since the legal action happens at the parts level and not at the whole level, let’s not worry about the whole and focus ourselves, our policies and our risk profiles on the parts.