Last Week in Tech Law & Policy Vol. 6: Self-Driven Tech

(by James Frazier, Student Attorney)

Planes, automobiles, and artificial intelligence. This week, the FAA released new rules regulating drones, Apple and Sony threw their hats into the autonomous automobile arena, and Great Britain announced that it will release new rules governing driverless car testing in the UK. Mike Hean, a Swiss entrepreneur, suggested that driverless cars should be programed with advanced artificial intelligence that would allow an Uber-style ride sharing app. Indeed, Hean believes that driverless car should be able to own themselves, and even earn their own income. Tying these threads together, MIT’s Tech Review released an article this week outlining some contemporary concerns about artificial intelligence.

Drones

This weekend, the U.S Federal Aviation Administration (FAA) released long-awaited rules regulating operation and licensing of small commercial drones. These rules include licensing requirements for commercial drone pilots such as passing a knowledge based license test, maintaining a valid drone operators license, and passing a TSA screening process. Commercial drone operators will have to register their drones with the FAA. However, the FAA is considering an exception for drones weighing less than 4.4 pounds. Although drone hobbyists will not have to obtain commercial or private pilots licenses, hobbyists will only be allowed to fly their drones during daytime hours, at speeds no more than 100 mph at no more than 500 feet.

Most interestingly, these new rules allow the autonomous flight of commercial drones, provided that the drone’s flight conforms with all of the regulations listed above. Commercial drone operators must also maintain a line of sight with their autonomous aircraft and be able to immediately recover manual control of their autonomous airplanes.

Driverless Cars

The FAA’s new rules could be be predictive of rules that will one day regulate driverless cars. This week, Apple announced development of an electric, and possibly driverless car. Although Apple is notoriously secretive about research and development, there’s a healthy amount of online speculation about the Apple car. Not to be outdone, Sony, one of the world’s largest image sensor developers acquired a 2% share in ZMP, a Japanese startup that makes robot cars. Sony and Apple join other industry giants such as Google, Mercedes-Benz, Audi, BMW, Nissan and Tesla in the driverless car development space. It’s becoming increasingly apparent that driverless cars will be a viable mode of transportation within the next 25-50 years as firms continue to enter this market space. But how will we regulate and license driverless vehicles?

Great Britain began answering that question this week when the UK announced its intention to become a world leader in testing and use of driverless cars. The British government will release a code of acceptable driverless car testing practices this spring. By 2017, the UK will do a total rewrite of its Highway Code in order to accommodate driverless cars on British roadways.

Although companies will be allowed to test driverless cars anywhere they wish within the UK, the UK’s new rules require human testers to be present in the driver’s seat at all times during testing. Further, human testers need to be able to take manual control of their self-driving cars at any time during driving. These rules are similar to the FAA’s recent drones regulations.

Artificial Intelligence

Yet some entrepreneurs like Mike Hean are not content with humans remaining in the driver’s seat. Indeed, Hean believes that driverless cars should be allowed to own themselves. But more than that, Hean believes that driverless cars should be able to pool their resources and operate an Uber or Lyft style ride-sharing app. This service would allow driverless cars to connect with pedestrians using a smartphone app similar to Uber’s. According to Hean, driverless cars could use the profit generated from this service to hire human programmers to perform upgrades on the auto’s software. Alternatively, autonomous autos could pool their profits and commission factories to build newer model driverless vehicles. Hean ultimately envisions a world where autonomous cars could emigrate from city to city to meet increasing consumer demand. Or, these freewheeling autos could voluntarily place themselves in long term parking, and power off until consumer demand required more autonomous cars on the road.

You’re not alone if thoughts of autonomous cars moving humans around major cities evokes images of Skynet or the Matrix. Hean acknowledges this fear in his article. Yet Hean indicates that his autonomous cars will not be so smart that they will push us towards the dreaded singularity. The singularity is the moment at which artificial intelligence will surpass human intelligence. Stephen Hawking and Elon Musk both view recent developments in artificial intelligence with trepidation—Musk, so much that he donated $10 million to an organization created to combat “existential threats to humanity.”

There are two noteworthy thought experiments on artificial intelligence. One asks what would happen if you gave a paperclip maximizer machine stellar artificial intelligence. And the other asks how well a human in windowless room could learn a foreign language. Early artificial intelligence researchers discussed educating machines in the same way that we educate children. At some point in their formal education, children begin to make value judgments. Machines can’t make value judgements, yet. But anything seems possible if they could.