The EU AI Act takes full effect on 2 August: what changes for maritime AI tools?
EU AI Act fully applies 2 August 2026 with €35M or 7% turnover penalties for prohibited maritime AI; €15M or 3% for high-risk systems.
IMPACT SNAPSHOT
Penalties reach €35M or 7% of global annual turnover for prohibited practices; €15M or 3% for high-risk system breaches — the cost of non-compliance is balance-sheet scale
The EU AI Act's high-risk obligations and enforcement powers apply from 2 August 2026; the law is horizontal across sectors (no maritime-specific exemption)
- Who is affected
- Worldwide
- When
- Full effect 2 August 2026 — inventory the risk class of the AI tools you use now
- What to do
- Clarify with your vendor which risk category your fleet's routing/compliance/planning AI falls into; require explainability and an audit trailNo maritime-specific enforcement case exists yet; this article addresses the general framework's likely effect on maritime AI tools — sector interpretation is not settled case law
For maritime companies putting AI tools into operation, a new compliance obligation is coming into focus: the EU Artificial Intelligence Act enters full application on 2 August 2026, together with its high-risk system obligations and enforcement powers.
The penalty framework is severe: for prohibited AI practices, fines can reach €35 million or 7% of the company’s global annual turnover. For breaches of high-risk system requirements, the ceiling is €15 million or 3% of total worldwide turnover. The law is horizontal; maritime has no exemption of its own.
The critical question for shipping is this: which risk class do the AI tools used across the fleet — route optimisation, compliance reporting, crew planning — fall into? That classification determines the transparency and documentation obligations the vendor and the user must meet.
Note: at the time of writing there is no maritime-specific enforcement case under the EU AI Act; the assessment here concerns the general framework’s likely effect on the sector.
Apeks view — “We use AI” is no longer marketing; it is a declaration of liability. The real question the law brings to shipping is not technical but governance-related: can you explain what your tool does, what data it relies on, and how it reaches its decision? A black box you cannot explain is a black box that carries penalty risk. What we have argued from the start — explainable, evidenced, contestable systems — stops being a matter of preference under regulation. When you choose your tool today, ask: can this system show its reasoning to an auditor?
Spot an error? Request a correction
Apeks Tech Editorial Team
Sourced curation
The Apeks Tech editorial team — sourced briefs and engineering-led curation; sources are listed on every brief. Editorial policy →
Forward this to a colleague
Should your DPA, technical superintendent or a fellow owner see this? Forwarding costs nothing.
Was this forwarded to you? Take your own copy → subscribe here.