ARTIFICIAL intelligence will revolutionise shipping, as long as it can be trusted.

But whether it can depends on greater understanding and robust safeguards that will not arrive soon if they arrive at all, leaving many unsure of what to do next.

And the potential drawbacks of each use — from falling foul of intellectual property, to the possible extinction of humanity — are leaving plenty of companies unsure how to proceed – and especially how much they should be spending.

Speakers at Tuesday’s Demystifying AI in Maritime conference said the industry was still experimenting with the number-crunching power of large language models.

Dr Amy Parkes, a machine learning and data science specialist at Arcsilea, a maritime data consultancy, said the two main questions from companies were “what system should I be using?” and “how much is it going to save me?”.

Parkes said this was particularly hard to assess for maritime companies, because of their diversity of ships, sizes and operations.

“We do exactly the same thing on two different ships, and the potential for savings is massively different,” she said.

Wallem Group IT director Steve Whitby said the technology had yielded plenty of quick wins already, from SIRE ship inspection reporting and safety assurance to master data management and consolidating financial data.

“There’s a huge amount, as you can probably guess, of accounting and spreadsheets and systems involved in looking after numerous shipowners, all with their different requirements,” he said. “And that’s made a huge difference there.”

Fred Olsen Cruise Lines is testing AI-enabled cameras above bins on its cruiseships, to gather data on what and how much food is wasted. But the data still needs to be checked manually.

Seafarers spoke of using ChatGPT as a therapist, teacher and digital assistant — as long as they have the bandwidth.

But pitfalls of AI are already emerging, on and offshore.

Maritime Passport co-founder Martin White spoke of the potential for digitising seafarer paperwork, creating a “trusted environment from the issuer to the receiver”.

But he added: “There is also a large problem within the industry of fraudulent documents, and AI is only encouraging that.”

Whitby said cyberattacks on Wallem had tripled in the past year, with 114,000 attacks in the past quarter alone. “We’re absolutely convinced it’s AI behind it,” he said.

Seafarers had reported the growth of an “always-on” culture on board, where greater ship connectivity also meant more texts and emails to reply to, and more admin rather than less.

Companies were urged to put their own policies in place and make sure they knew what problem they wanted AI to solve before rushing in.

“There’s no point investing in any technology until you have your house in order,” Parkes said.