Seatrade Maritime is part of the Informa Markets Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

AI and chartering – shipping is the nemesis of algorithms

Photo: YouTube screenshot autonomous car shut down with traffic cone
SSY Global Head of Research Roar Adland likens algorithms difficulty of dealing with the unpredictability of shipping to that of disabling an autonomous vehicle with a traffic cone.

Speaking at the Singapore Maritime Week Conference Adland said he believed the impact of AI on conducting business in shipping would not be that great compared to sectors with high frequency trading such as shares.

“I think it’s fair to say shipping is the nemesis of algorithms, it’s really everything algorithms are not particularly good at,” he stated.

Adland said that algorithms are good at sifting through data sets and the categorisation of low dimensional outcomes. “They are good at supporting decision making in high frequency businesses like share trading for example. They are also good at operating in clean virtual environments and shipping is not that,” he explained.

While shipping has a lot a data the informational content of that data may not actually be that large or useful. Adland gave the example of AIS data showing a map of the global fleet that comes from some 2 billion data points a day. However, this data does not change very much. “If you take a snapshot of this data, and you do that once an hour, you will have 24 pictures in a day, which look very similar,” he said.

Another issue ia that while there is public data such as AIS that everyone can access much of shipping remains highly private. “It’s also an industry that is dominated by private data – so there is data that only you know, or your closest colleagues and it’s very hard to access that data across the entire industry.”

The issues with data then combine with shipping being a low frequency business, and having non-standard contracts, with no two charter contracts being exactly the same.

On top of this is the is that when things go wrong in shipping it is not necessarily in a way that’s predictable. “Things in shipping go pear shaped all the time but they don’t to go pear shaped the same way they did previously.”

So, the lack of predictability and complex operating environment was not one which was suited to AI algorithms. Adland said the situation with shipping was similar to that of autonomous vehicle which uses terabytes of data to operate in the most efficient yet easiest way to disable is place a traffic on the car’s bonnet confusing its sensors and causing it to shut down.