AI Becomes Prevalent in Maritime But Needs Greater Governance and Verification

Artificial intelligence is becoming more embedded in maritime administration and operations, but shipping companies need to invest in more governance and encourage greater verification of results to raise trust in AI outputs, according to new research. Research by Thetius and Marcura found that around 63 percent of maritime professionals use AI every day, but just 8 percent describe their organisation as mature and governed in how it uses these programs. The findings point to a gap between adoption and oversight.

 

The Growing Adoption

 

AI is increasingly embedded in maritime. It features in administration and operations. Companies see growing benefits. Employees also recognise its value. This drives rising adoption.

The research quantifies this adoption. Around 63 percent use AI daily. This reflects widespread uptake. Yet governance lags behind. Just 8 percent describe mature governance.

 

The Governance Gap

 

A clear governance gap exists. Adoption has outpaced oversight. More verification is needed. Clear ownership is also required. Governance must improve for full trust.

The gap has practical implications. Most respondents felt data was not ready. Over 90 percent held this view. This concerns AI reliance. It limits trust in outputs.

 

The Verification Burden

 

Verification currently takes significant time. Around 55 percent spend an hour weekly checking. Some spend much longer. This can reach five hours or more. This reflects a verification burden.

Yet AI still saves time overall. Most users reported net time savings. This holds even with checking. Most would check important recommendations. A named person should remain accountable.

 

Read More: https://oceaneconomist.com/articles/zodiac-maritime-lng-hanwha-order

 

The Data Challenge

 

The research identified a data challenge. Maritime does not lack data. The problem is fragmentation. Context is scattered across sources. These span contracts and systems.

Company leadership framed this issue. Putting AI on fragmentation does not help. The opportunity is bringing context together. This occurs at the decision point. It then learns from outcomes.

 

The Path to Trust

 

The research points to earning trust. Human oversight remains important. However, it should be proportional. This reduces unnecessary checking. Confidence should grow over time.

Trust depends on transparency. Systems should show their basis. They should be tested against outcomes. They should improve from corrections. This earns trust rather than asking for it.

 

Significance of the Findings

 

The findings offer clear guidance. Companies should measure the full process. This covers the human-AI combination. They should differentiate verification types. Necessary checking differs from uncertainty-driven checking.

The findings support AI maturity. Human corrections should improve performance. This builds future reliability. Reliable outputs reduce checking. This advances trust in maritime AI.