newswire: 8/18/2020
- With Japan’s population aging fast, the country’s shipping industry is pouring resources into developing autonomous ships. Like driverless cars, driverless ships pose plenty of challenges—but with more than half of the industry’s mariners over age 50, shipping companies need to innovate. (The Economist)
- NH: We have often discussed Japan’s shrinking population and the prospects of automation solving future worker shortages. (See "Low Fertility Countries Look to Robot Soldiers.") Here is just another example. A trading nation that is itself composed of over 400 islands, Japan has always depended critically on ocean transport--yet half of Japanese mariners are over the age of 50. Question: Who will steer the ships in thirty years? Possible answer: self-driving ships.
- This idea is not far fetched. And it's actually not that different from how airplanes are now flown. For years, commercial aircraft have been equipped with autopilot sensors that detect weather patterns and air pressure changes. The autopilot is so advanced that newer planes today routinely land themselves.
- Autonomous ships would similarly be outfitted with sensors to detect water currents and weather patterns. The benefit for Japan's maritime industry and hence its overall economy would be that these ships wouldn’t need large crews. Currently, cargo ships are typically manned by 20-30 workers who check systems for minor repairs. With an autonomous ship, you could cut the crew size down to 3 or 4 people. Their only job would be to manually override the system if something went really wrong.
- In certain situations, autonomous ships could be a real improvement over manned ships. Consider piracy. When they take over an autonomous ship, the pirates would have to figure out how to hack into the ship's electronic AI. Any remaining crew could retreat to a safe room. Control meanwhile would revert to a Japanese-based HQ. It's not a perfect solution, but it presents fewer vulnerabilities than today.
- The author of The Economist article doesn't think driverless ships will happen any time in the near future--and at one point implies that automating ships is more difficult than automating cars. I have to disagree. As I have argued elsewhere (see "Driverless Cars: Unsafe at Any Speed?" and "Have Autonomous Vehicles Hit a Roadblock?"), level-five automation for cars driven in bustling urban traffic is probably the very hardest of all robotic tasks to master. And it will likely be one of the very last to be accomplished. Ships, unlike planes, don't require split-second execution tolerances. And neither ships nor planes require what automobiles require, and that is constant availability of higher-order human intelligence for interpretation and intervention.
