When the robots take over

Imagine a delivery truck. It belongs to a contract delivery company. It rolls down the interstate to make a delivery. It has a really excellent driver, who happens to be a computer.

The computer drives day or night, eyes never closing or wandering for an instant, mind never drifting, running an operating system honed first by millions of real and simulated hours on ice, snow, oil, sand–it has experienced every one of these conditions and a hundred others, thousands of times. A distributed network of inexpensive sensors and embedded computers in each of its major components gives it access to real-time information about both its internal state and the external environment; including a complete spatial representation of the road built in multiple layers of abstraction, from downloaded maps of the upcoming road to real-time reports of traffic conditions, weather, construction, etc.

Much of the information comes from vehicles ahead on the road; the driver is in turn passing its own data back to other connected vehicles on the same network (the Interstatenet?), which also enables a broadband internet connection to be shared among all the vehicles. This significantly increases the range of the private- and government-owned wireless networks by rebroadcasting the signal from vehicle to vehicle (which requires trivial amounts of power in comparison to, say, air conditioning, and has the advantage of increasing the available bandwidth in direct proportion to the number of cars on the road, which is a convenient proxy for the load on the network. Elegance).

The truck can “see” in several spectra via any combination of near- and far-infrared cameras; laser, radar or sonar location; and visible-light cameras. This provides excellent night vision and, critically, long-range detection of warm, moving objects. Better informed than any human driver, it also has reaction times measured in milliseconds.

It might be as “smart” as a dog or a mule, but probably dumber–more like a reptile. That’s probably about all it would take to take to run a body as complicated as a truck safely from point A to point B while obeying well-programmed safety rules, avoiding collisions and cargo damage. It could be made smarter, especially if it needed to deal with people a lot, but it probably wouldn’t need to be–because it would mostly be interacting with other robots. The truck has no human on board. When it arrives at its destination around 2 am, it will back accurately up to an automated loading dock, which will digitally verify its identity and sign a contract verifying receipt, as authorized by the business owner earlier–or perhaps approved automatically by an AI “employee” responsible for managing inventory. Robotic hydraulic lifts will roll in and remove the palletized goods under direction from the AI manager, perhaps for inspection later by human employees–or their artificial equivalents.

None of this need be supervised by a physically present human, although internet-connected security cameras are so cheap and ubiquitous that it would be trivial for an absent owner or manager to keep a tight watch regardless. In practice, this will become increasingly unnecessary–and a good manager will likely learn to delegate as much or more to a reliable AI overseer as she would to a human. AIs would be commercial products just as software is now, and no company would survive if there was any chance their AI might embezzle you or hand over trade secrets for bribes. AIs won’t generally be motivated by money anyway–it would be counterproductive to program or select them (evolutionary algorithms having taken over many of the duties of today’s designers and engineers) for that trait, rather than for complete loyalty to their owners and/or the public good.

Wait a minute–no people? So this one dude could run a whole, real-life business (or, for that matter, a robot or AI could run one), and not have any human employees at all? Sounds great, right? No payroll! No crazy employees to fire!

Wait. Wouldn’t that be, well, kind of a disaster, if every business started doing it? Won’t millions of truck drivers and dock loaders and airline pilots and bartenders and bridge architects and… be out of their jobs? AIs and robots can probably be designed that will eventually be able to perform almost any human occupation, without needing to be paid any money, sleep, or go home to their spouses. Humans won’t be able to compete at all!

That’s right. We probably won’t. Which is why we’ll need to learn to play a different game. Most people won’t have “jobs” in the current traditional sense, because the majority of them will eventually be done by machines. On the other hand, material productivity will be skyrocketing, with fewer and fewer human-hours of work required for anything we regularly produce or consume.

This could have a few outcomes, and these depend heavily on what measures we take to remedy this “problem” before it runs too far. If left unchecked, too many people unable to buy an increasing amount of cheaper and cheaper goods is the kind of runaway situation that leads to traumatic “market corrections” as seen in depressions and recessions the world over.

Alternatively, there are several steps we could take to improve the social safety net and provide basic economic security to all Americans (eventually, all the world! Huzzah!). Such measures include increased progressivization of taxes, a land or capital tax, a basic income, and true universal health coverage. The general idea is to move just enough money to the hands of people who need to buy goods, at such a level that everyone is reasonably certain to have their material needs met, whether they earn income at a traditional job or not–thus redistributing enough of the wealth generated by our automated servants to continue greasing the economic wheels for the foreseeable future.

In theory, this should have several positive results, including the wholesale elimination of abject poverty. Human beings would also be largely freed from menial, dangerous and unfulfilling jobs. Most of the evidence I have read (and experienced) indicates that people are much more likely to produce their best, most creative work when free from serious worry about their next meal, their children’s health, etc, all of which create a sort of mental “overhead cost”, which lowers effective IQ and impairs impulse control. Let me stress that this is a completely uncontrollable response of the human animal to an environment which is causing serious psychosocial stress in the form of resource scarcity (jobs, money, respect, dignity) and entrenched social norms of inequality and injustice. It has little or nothing to do with one’s innate intelligence, worth or talents, but is akin to being dragged down by mental parasites, which divert resources of thought and creativity from one’s other needs. This suggests that a widespread alleviation of the mental burden of extreme poverty through direct financial support of all citizens’ basic needs might unleash a great deal of currently-underutilized human creative potential.

The barriers to small-scale entrepreneurship (“lifestyle businesses”) would similarly also drop dramatically, especially in a legal future where basic access to health care has been successfully decoupled from employment and income–a monumental task ObamaCare has recently nudged us a few inches closer to finishing.

So, a world where most people don’t need to “go to work” the way we understand it today, and certainly not full-time; where those who are required to do the few “necessary” jobs that have not yet been automated away will generally be well-paid and well-respected for doing so; where many more people are economically free to pursue their chosen creative expressions or life vocations, without fear of bankruptcy, starvation or unexpected illness; where many of these people will go on to make lots of money. Where do I sign up?

When the robots take over, it won’t be to turn you into a human battery for their electricity farms (a thermodynamically impossible proposition anyhow, but we’ll leave the physics of The Matrix for another time). It will be to take your hands off the steering wheel and onto your hobbies, your attention off your finances and onto your passions. That is, it could be–if we react to this onrushing wall of change with openness and a positive spirit of adaptation, rather than fear and reactionary resistance.

2 thoughts on “When the robots take over

  1. Tim

    Reminds me of a season 1 episode of Star Trek: TNG where because the generation of resources and energy is so efficient scarcity is largely a thing of the past hence classical economics is also antiquated and obsolete, Picard explains to cryogenically frozen (it was the late 80s after all) and thawed people from the 20th century that people now work for the betterment of themselves and mankind not for any other purpose.

    1. ACH Post author

      Cool. As usual, Star Trek was ahead of its time.

      The interesting part to me is that we’re already practically there in the more developed countries. The main economic challenge we have in the US is income inequality, which could be addressed with a combination of measures given the political will. With a median household income in the $40-50K range, actual material scarcity is a non-issue (never mind that many people mistake a perceived lack of consumer luxuries for true physical deprivation, i.e. starvation or lack of shelter–and we don’t do enough to help the small fraction that actually experience the latter, despite the negligible cost compared to, say, bombing Syria again).


Leave a Reply

Your email address will not be published. Required fields are marked *