What’s Next for Technology and How it Could Benefit the U.S. Economy

For the last two hundred fifty years, new technologies have driven the growth of the US economy. During the Great Recession, concern has developed among economists that the rate of technological innovation in the future will not stimulate enough economic activity to grow the economy at an average rate of 2.5% or better (for example, see Robert J. Gordon “Is U.S. Economic Growth Over?”).

I am not an economist and I don’t know the answer to the question. However, as a technologist,  I do know that the U.S. is on the brink of an interlocking series of new technologies which will completely reshape the economy and society over the next twenty years. Over the next few weeks, I’ll discuss five emerging technological applications that will have a major impact on our economy and society. Each is driven by the rise of the Internet, the solution to a set of “big problems” in the field of Artificial Intelligence (AI) or both.

Artificial Intelligence and Robotics

Thirty years ago when I was starting out as a graduate student in computer science, AI had four “big problems” it needed to solve, and we were not sure there was a solution:
  1. Computer vision and environment sensing – making it possible for computers to see and use other sensors to understand its physical environment.
  2. Computer speech recognition and natural language processing – making it possible for computers understand human speech and understand human languages
  3. Machine learning – computers that learn from experience, particularly pattern matching
  4. Computer manipulation and navigation – basically robotics.   The ability to navigate and manipulate the physical world.
Today, practical solutions and practical applications exist for all of these areas. We can create robots that can see and sense better than humans, that can understand human speech and language well enough to perform many tasks, that can learn and can navigate and manipulate the physical world.

What have been particularly powerful are robots that in addition to the above, communicate via the Internet.   Communication with the Internet first gives a robot access to much greater information and processing power. Second, it makes possible the model of the “semi-autonomous robot”, a robot that is remotely controlled by a human. This allows human judgment to control a robot. Human judgment is vital and not yet, if ever, reproducible by a machine. An example of a semi-autonomous system with which you might be familiar is the battle drone: a flying robot remotely controlled and capable of the sort of reconnaissance and attack that used to require a manned aircraft.

First Example: Driverless Cars are on the Horizon

One practical example of a semi-autonomous robot that is practical given today’s technology, which will likely be deployed widely in the next ten years and have an enormous impact on the economy, society and your life is the driverless car. If you are not familiar with driveless cars, I recommend this quick article from the New York Times: “Yes, Driverless Cars Know the Way to San Jose”.

One of Google's prototype driverless cars. A Lexus hybrid.
We can be confident that driverless cars will be deployed widely over the next ten years because they have the following qualities:
  1. No technology break-throughs are necessary for widespread deployment; prototypes are deployed.   All necessary technology already exists and only step-wise refinement and economies-of-scale are necessary to reduce costs for consumers.
  2. There are significant advantages to individual consumers for the new technology that would justify the price.
  3. The technology is charismatic and a status symbol for early adopters.
  4. The technology has significant advantages to society.
  5. There is a step-wise path for deploying the technology in small increments.
Let’s look at the step-wise path for deployment.   High-end cars already have a variety of sensors that can tell how fast this vehicle is going and where other cars are.   This allows a “smart cruise control” which makes accurate judgments about when it is necessary to slow down in response to other vehicles.

Likewise, these cars can parallel park themselves using sensors. It is likely that more and more well-defined driving tasks can be taken over by the car. The example given in the article is driving in a traffic jam. As more cars include more of these features, society and consumers will become more comfortable with the notion that the car drives itself. Eventualy, the fully driverless car is possible. Clearly these auto-driver features are charismatic because luxury car marketers are already using them to sell.

For an individual consumer, the benefits are enormous. Driverless cars have lower operating costs because they can make better fuel-efficiency choices. They are much safer, because most car accidents are caused by human error. They can safely drive faster because the car can react more quickly to changing circumstances than a human driver. A driverless car allows people to use trip time to be productive or entertaining because it is not necessary to pay attention to the road.   Parking in cities becomes easier because your car can drop you off, drive to a parking location and drive back to pick  you up later. In fact, it would be quite possible to not own a car, but rather subscribe to a driverless car service which guarantees you will have a fleet car at your disposal when needed.

For society, the benefits are enormous: less congestion, less parking necessary, greater productivity and most of all, fewer deaths and injuries.   Each year, about 5.25 million reportable accidents occur in the US. In 2010, 32,885 people died in auto accidents.   Hundreds of thousands were injured.   Driverless cars will have an enormous economic and social impact.
GWI Blog