Photo credit: Pexels

Amber Foster, co-founder of Human Superpowers and former-chair of Shoreham Port Board, outlines how board members in maritime organisations should approach AI implementation in order to maximise impact.

Across every sector, boards are asking the same questions: “What should we do with AI?”, “How can we go faster, where can we save costs”. I hear many variations of these, but boards are missing the big vital question: “what could AI help us do that was impossible before?”

Maritime is no different to many other sectors. Most boards have mandated the same setup, or at least have allowed it to exist. A responsible use policy (tick), an AI champions group (tick). Pilots running left, right and centre (tick). Unloved prompt libraries (tick) and adoption metrics that show initial curiosity, dropping away to disillusionment (tick). And a growing sense across the organisation that people just don’t get this AI thing.

And that’s because the board’s ambition is too small.

What AI ambition looks like

In January 2026, The Lancet published the results of the MASAI trial — a Swedish study tracking over 105,000 women through AI-supported mammogram screening. The researchers weren’t asking how AI can make us faster or more efficient. They asked: how can AI help us find cancers we’re currently missing?

The result: 29% more cancers detected. Fewer aggressive and advanced-stage diagnoses. No increase in false positives. And a 44% reduction in radiologist workload — but that wasn’t the goal. It was a side effect of aiming at the right thing. And interestingly, AI plus human judgement turned out to be the best combination

Maritime boards should be asking the same kind of question. What would it look like to be safer than you’ve ever been? What patterns are sitting in decades of incident data, maintenance logs and near-miss reports that no human team has the capacity to read? Every port and shipping operator is sitting on operational data that could change how they manage risk — if they asked the right question of it.

The tool addiction

In most cases, I see boards obsessing over tools and licences, often driven by FOMO and competitor press releases. There’s an addiction to buying tools (or at least trialling them) before anyone has asked what problem they’re solving.

If your AI ambition starts and ends with “go faster”, “reduce costs”, “let’s do AI ’cos everyone else is, then you’re chasing short-term wins. You’ll speed up what you already do and you won’t discover what you could do differently. Vendors know this — they’re selling solutions to problems boards haven’t properly defined, and it’s driving a lot of AI disillusionment.

Boards need to ask some uncomfortable questions. Most organisations’ data infrastructure was built long before AI was a realistic tool, so gaps are expected — but the board needs to know what state their data is actually in. Do you understand enough about the underlying technology to challenge what vendors are telling you? And have you set a goal that goes beyond being more efficient?

The governance vacuum

If the board doesn’t set a clear direction, someone else will fill the gap, and it’s usually the AI governance and compliance teams. Their natural skew is to default to risk management over ambition. 

That’s not their fault – they’re doing what they’re trained and paid to do, which is protect the organisation. Without a clear strategic goal and defined risk appetite from the board, they default to caution and everything gets treated as high risk. 

I’ve seen organisations requiring 100-question in-depth questionnaires for low-risk AI use cases — the same form whether you’re deploying a chatbot to answer FAQs or an AI system making safety-critical decisions. Boards need to create risk-based governance channels – quicker processes for low risk tools, space for safe experimentation and deep dives into risk where it’s proportionate. One size governance does not fit all.

And they need to resource committees properly. Right now, this work gets added to the busy privacy, info sec or legal professionals’ day jobs. People get put on these committees because they’re ‘good at compliance’, not because they understand the technology. Governance teams need training, expert advice on hard concepts like bias, and they need to accept that as soon as policy is written, the tech has already moved on. They need to be comfortable advising on the fly as new technology gets thrown at them.

I helped one organisation with their policy on AI transcribing tools, within days wearable tech that records meetings was all over TikTok and the policy immediately out of date.  

Many organisations are winging it or letting compliance fill the vacuum. They’re picking up what worked for other tech implementations or privacy governance and applying it to AI. But AI has changed the game and needs governance that can move at the speed the technology is changing.

The answer is governance that’s proportionate to actual risk – rigorous where it matters, lighter where it doesn’t. But that only works when the board has set the strategic goal and defined how much risk the organisation is willing to take. Without that, governance has nothing to calibrate against.

Three questions boards should ask every quarter

Q1: What would we do that’s currently impossible without AI?

Something you genuinely cannot do today. If the answer is “nothing”, your AI ambition is too small. As the technology develops at speed, you’ll need to reset this goal often, but aiming high means you’re building towards something that matters – not just going faster.

Q2: What value did we unlock with our AI initiatives last quarter?

“80% of people used AI last quarter” is a useless metric and so is “we trialled eight new products”. Ask for data and stories about what value was realised – for the organisation, its people, its stakeholders and its mission. And give explicit permission to fail. If you trial something, see no value then move on. And right now, think carefully about signing multi-year deals for tech that’s improving this rapidly.

Q3: Have we given our governance team clear direction on risk appetite?

If you haven’t told your governance team what level of risk you’re comfortable with, they’ll default to maximum caution. That’s rational behaviour from them – and a failure of direction from the board.

The job is to set the ambition, define how much risk the organisation is willing to take, and make sure governance serves your strategy. But you do need to have experienced what AI can actually do, even at a basic level. If you haven’t felt what’s possible, you’ll keep setting goals that are too small.

———————————

Click here to sign up for our free monthly ‘De-Mystifying Digitalisation and AI in Maritime’ newsletter and get the latest cross-sector expert insights, news, interviews, videos and case studies straight to your inbox.