AI: Changing the Game — Today and Tomorrow
Where artificial intelligence is genuinely changing maritime operations today, and where it is heading next — told as a journey from the 'dark ages' of vessel performance to a practical, incremental path toward autonomous, efficient sailing.
A conference talk on the real state of AI in shipping — the recorded version of the talk later given at COP29 in Baku. It traces the industry across three stages: where we were, where we are now, and where we're heading.
What the talk covers
- The "dark ages" of vessel performance: charter-party speed/consumption tables built from memory and guesswork, with little visibility into fuel, fouling, or weather impact.
- "You can't improve what you can't measure": high-frequency data solved the frequency problem, but data accuracy — not just volume — is what makes AI trustworthy.
- Where we are today: continuous fouling assessment, optimal routing, main-engine fault detection, and retrofit / hull-cleaning timing — a shift from uncertainty to control.
- Top-down vs bottom-up: clean-sheet concept ships (e.g. Wallenius Wilhelmsen's Orcelle) versus incremental automation of real daily problems — like cars going from manual to cruise control to near-self-driving. Bottom-up solves today's problems now.
- Proof in progress: a "cruise control" system that loads a voyage's optimal speed profile and commands the main engine directly; and a hybrid physics-plus-data ship-motion model shown berthing autonomously in a simulated real Japanese port.