Like the self-driving vehicles, the legitimate challenge is — who will be responsible in case of a collision?
By Kiran N. Kumar
Recently, IBM’s autonomous ship named Mayflower 400 tried to follow the same route of the historic Mayflower ship that sailed 400 years ago in August of 1620. Bereft of any crew, it sailed successfully from England to America, aided entirely by Artificial Intelligence (AI).
The ship’s 40-day journey heralded a new beginning for AI-controlled self-navigating ships on the lines of self-driving vehicles. Flagged off from Plymouth, UK on April 29, the ship reached the coast off Northern America in Halifax, Nova Scotia on June 5.
Equipped with six AI-powered cameras, 30 sensors and 15 Edge devices, the ship’s AI captain could adhere to maritime law while making crucial split-second decisions, like rerouting itself around hazards or marine animals, all without human interaction or intervention, according to IBM.
A first-of-its-kind, Mayflower 400 is a joint project of IBM and ProMare and it had a trained “AI Captain” or AI fed with if/then data to make real-time rules on its own. It could learn “from the outcomes of its decisions, makes predictions about the future, manages risks, and refines its knowledge through experience,” IBM stated.
However, the ship, destined originally to reach Washington, DC, ended up at the Canadian port of Halifax due to weather conditions. Undaunted, the engineers are preparing for another voyage overcoming the challenges.
Challenges ahead for autonomous ships
According to a study by insurer Allianz in 2018, human error was behind 2,712 casualties in 2018, costing $1.6 billion in losses from 2011 to 2016, and 56% of them were cargo ships.
And more than 75% of such maritime accidents are attributed to fatigue, judgement errors, negligence and inadequate training.
From the cost point of view, a study by the Technical University of Denmark showed that an onboard crew member costs about $150,000 annually, or 30% of the total cost for a cargo ship goes into crew costs.
The facilities required on board for the crew will give way to more space and reduce the weight and fuel consumption by autonomous ships. At the same time, people replaced by autonomous ships can be employed onshore for control and remote operations.
Some have argued that autonomous ships can be built to make them free from pirates as access to cargo can be made unavailable as control centers can immobilize the ship. With no crew, hostage and ransom demands prove futile.
But what about SoS? When someone requires rescue under Rule 5 of the International Regulations for Preventing Collisions at Sea (COLREG) or under the International Convention for the Safety of Life at Sea (SOLAS), it remains to be seen how autonomous ships should react or help in rescue operations, such as picking up survivors in case of a shipwreck.
While the International Maritime Organization (IMO) has already begun review of certain portions of treatises that affect autonomous shipping, there are states which developed and implemented their national regulation as Russia has done.
Read: The Robot Ships Are Coming … Eventually (October 30, 2020)
But the threat of cyber attacks looms large, snapping communications with jammers or alter GPS signals. What happens when the ship’s data gets compromised? As the self-driving vehicles are already facing, the legitimate challenge is — who will be responsible in case of a collision?
As several parties from the company that is operating, the software provider, hardware provider and the onshore monitoring stations are involved, the same regulatory framework that is emerging for self-driving cars may apply here soon, though it is still in its nascent stage.