Considering the efficiency challenges ports face and the periodical/recurring nature of tasks within these transportation hubs, automating them might seem like a good idea.[1]Chu, Fox, Sven Gailus, Lisa Liu, and Liumin Ni. “The Future of Port Automation | McKinsey,” December 4, 2018. … Continue reading Data-driven solutions to crane operations, arrival, container transport from ships to trucks and from trucks to ships, and many other processes promise increased productivity and profits to port owners on the surface. However, initial capital and social costs render full automation an undesirable option. Automating part of operations emerges as a more pragmatic and ethical solution.
Port owners would need to heavily invest in data collection and organization, Information Technologies (IT) and Artificial Intelligence (AI) specialists in particular, and updated physical layout that would facilitate a newly automated transportation system. Five main categories of automation will all need to be built from the ground up: digital infrastructure (gates, ground transport, etc), soft- and hardware to control such infrastructure, the computational decision-making “brain”, standardized and real-time data transfer, and finally human user interfaces.[2]Chu, Fox, Sven Gailus, Lisa Liu, and Liumin Ni. “The Future of Port Automation | McKinsey,” December 4, 2018. … Continue reading
A large port would require about $300 to $500 million to fully upgrade, with most of the cost coming from automated infrastructure and a sizable part from training the personnel.[3]Weeks, Kelly, Purnendu Mandal, and Kabir Sen. “Advancements in Technology and Potential Impacts on Port Automations Decisions: The Case of Port of Singapore.” In Entrepreneurship in Technology … Continue reading However, the return on such an investment so far was lackluster. Ports or their individual terminals that did undergo automation report a decrease in productivity of about 7 to 15 percent, mostly due to increased delay during exception scenarios.[4]Chu, Fox, Sven Gailus, Lisa Liu, and Liumin Ni. “The Future of Port Automation | McKinsey,” December 4, 2018. … Continue reading Due to the lack of standardized process[5]Chu, Fox, Sven Gailus, Lisa Liu, and Liumin Ni. “The Future of Port Automation | McKinsey,” December 4, 2018. … Continue reading and adaptability in algorithms (compared with more adaptable humans), completely automated ports average about ten moves per hour per crane less than non-automated ones.[6]Chu, Fox, Sven Gailus, Lisa Liu, and Liumin Ni. “The Future of Port Automation | McKinsey,” December 4, 2018. … Continue reading Because partial automation would not displace all humans on site, ports would not face such losses in productivity in that scenario. Until the development of super-AI, the combination of humans (to deal with exceptions) with some automation (to quicken more routine parts of container transport) will create the most optimal mode of organization for the ports.
Furthermore, automating routine jobs now done by humans is on its own ethically questionable. In the case of the Port of Singapore, there was not a significant net drop of employed people, but most of the employees had to switch their jobs to more skill-demanding ones, which they are not trained for, or were replaced by white-collar specialists.[7]Weeks, Kelly, Purnendu Mandal, and Kabir Sen. “Advancements in Technology and Potential Impacts on Port Automations Decisions: The Case of Port of Singapore.” In Entrepreneurship in Technology … Continue reading Automation has divided workers into the ones who can adapt on the go, and the ones who can not due to their background, which presents an equity challenge. Similar tendency might be expected in other developed countries like the USA, which is our focus.
One way to avoid this pressure on current port workers, while reaping the benefits of automation is doing so partially. If AI and its maintenance employees would only be used in parts of the process that are not done at all, such as virtual arrival, then no current workers would be displaced. However, through reduced idle time for ships and trucks efficiency would still increase, and new IT jobs would be created to support this digital infrastructure. Virtual arrival of ships, for example, is a new optimization technology that does not have existing humans working on it. Quay cranes, on the contrary, already provide employment and exceptional cases in crane operation would directly harm productivity if not resolved. Hence, it is not the best candidate for automation. Moreover, there should be less up-front costs in upgrading parts of the system, not its entirety.
References
