Artificial Intelligence (AI) is coming, and it's going to be devastating for some people. Think about this: why wouldn't the owner of a trucking company replace every truck driver with a safer, cheaper, and more reliable AI commanding the wheel? In this case like so many others, there really isn't an argument to be made to keep the humans, those individuals with families to support, children to feed and educate, mortgages and bills to pay to keep the roof up and the lights on. That's what's coming in the foreseeable future, and it's scary. In some ways, it's a typical technological disruption. The New York Times used to have a corps of typesetters, but technological innovation replaced them. The same happened with telephone operators, factory workers, and a lot of accountants. Algorithms replace algorithmic labor, as a rule.
AI aims to transcend the typical paradigms of algorithmic activity and include what was until now a purely human factor: creativity. And it's doing so by capturing and analyzing tremendous amounts of data. At it's essence, it's a very human thing to do. How else do we operate in our homes and our workplaces? We use our five physical senses, in varying capacities, to capture immense amounts of data, analyze that data using our brains, and then behave in a manner to cause an effect in our physical environment. That precisely becomes the role of AI in the fast-approaching future. The human has been broken down to his algorithmic form and has been rebuilt inside a silicon chip.
From one perspective, all human actions ultimately exploit the muscles and the brain. We manipulate our physical world using our muscles and control the mode and magnitude of that manipulation with our brain. Technological revolutions to this point have largely revolved around replacing the need for human muscles. Bulldozers, welding robots, and sewing machines are examples of such disruption. Muscular dexterity has been trivialized in the past, but the resulting economies were dominated by the mentally dexterous. The economic potential of mental dexterity, the only thing still exclusively human, is now at risk of robotic appropriation; many production functions that depended on "thinking", data analysis at its core, are liable to be replaced by stronger data analysis engines, AI equipped robots, muscle and brain.
In this new order, which forms of mental dexterity are ripe for relegation to irrelevance? My theory boils down to classifying the type of data needed to perform the mental acrobatics associated with any job, any production function. First, there is clean data, data that is easily quantifiable and highly correlatable with clear conclusions. This type of data is easy to capture and can be used to make statistically significant statements with the use of a relatively low number of variables. Then there is fuzzy data. This data is difficult to capture quantitatively, and furthermore is not highly correlatable or statistically significant without the explicit analysis of an enormous number of variables. To look at physicians, a radiologist may be said to work with clean data, as he looks at quantifiable images. Quantified deviations from the healthy case can be diagnosed. A psychiatrist, on the other hand, works with fuzzy data. Every patient's baseline healthy state is different, and deviations from the nebulous baseline tell the story. The patient might lie or tell a half truth. A patient's answer depends on the incredibly nuanced medium of human communication in which words, verbal tone, and bodily demeanor must be considered to make actionable conclusions. A soldier in a war zone may be working with clean data, in which the target and the enemy are clearly defined, but a security guard in a crowded mall works with fuzzy data, in which the target is tough to identify, and the mode and magnitude of action must be chosen based on a plethora of fuzzy variables. All of this means that, if you want to keep your job for the next 50 years, choose a production function which depends on fuzzy data. Or be that guy and create AI; be the disruptor instead of the disrupted. If you can't do either of those, then be prepared to say hello to your new robot overlords.