Control over AI
Blog
Compliance and AI 9 min read

The Netherlands has its NIS2 law: what applies from 15 August

From 15 August 2026 the Dutch NIS2 law imposes duties of care, reporting and registration on more than 8,000 organisations. The timeline and what to do now.

Abstract ascending blocks in teal
Quick answer

The Cyberbeveiligingswet is the Dutch transposition of the European NIS2 directive and applies from 15 August 2026. More than 8,000 organisations across eighteen sectors are subject to three duties: a duty of care (demonstrably manage cyber risks with appropriate measures), a reporting duty (an early warning within 24 hours and a full report within 72 hours of a significant incident) and a registration duty (register through mijn.ncsc.nl). The board must approve the measures itself and oversee them, and can be held liable for serious negligence. Maximum fines reach €10 million or 2 per cent of global annual turnover.

01

The Cyberbeveiligingswet applies from 15 August 2026 and replaces the earlier Wbni

02

More than 8,000 organisations across eighteen sectors fall under the law

03

Three duties: duty of care, reporting duty (24 and 72 hours), registration duty

04

The board approves the measures, oversees them and is accountable

05

AI tools are part of your attack surface and supply chain, and therefore fall under your duty of care

Almost two years later than intended, but it is here: on 7 July 2026 the Dutch Senate adopted the Cyberbeveiligingswet, the Dutch Cybersecurity Act, and on 15 August 2026 it takes effect. That turns NIS2 in the Netherlands from a directive you prepare for into a law you comply with. For more than 8,000 organisations, cybersecurity becomes a legal duty, with a regulator, reporting deadlines and fines.

From European directive to Dutch law

NIS2 is the European directive that raises the bar on cybersecurity well beyond its predecessor. A directive only takes effect through national legislation, so the Netherlands had to transpose it. That should have happened by 17 October 2024; in the end it took until August 2026. The Cyberbeveiligingswet replaces the old Wbni (the Dutch Network and Information Systems Security Act) and passed parliament together with the law on the resilience of critical entities.

  1. The NIS2 directive enters into force

    The successor to the first NIS directive becomes European law. Member states have until October 2024 to transpose it into national legislation.

  2. European transposition deadline

    The Netherlands misses the deadline: the old Wbni stays in place longer than planned. Registering with the NCSC is possible on a voluntary basis from this point.

  3. The Dutch Senate adopts the Cyberbeveiligingswet

    Together with the law on the resilience of critical entities. That makes the start date final.

  4. New

    The Cyberbeveiligingswet takes effect

    The duty of care, the reporting duty and the registration duty apply to more than 8,000 organisations across eighteen sectors. Registration runs through mijn.ncsc.nl.

The Cyberbeveiligingswet is the Dutch transposition of the NIS2 directive and replaces the Wbni.

Who is covered

The law covers organisations in eighteen sectors, from energy, drinking water and transport to healthcare, digital infrastructure and government. Within those sectors a size test applies: mid-sized and large organisations are covered, as a rule of thumb from roughly fifty employees or ten million euros in annual turnover. Some organisations, such as providers of essential digital services, are covered regardless of size.

The law distinguishes between essential and important entities. The duties are largely the same; the difference lies mainly in supervision (proactive for essential entities, after the fact for important ones) and in the fine ceilings. If you are unsure of your status, the FAQ from the regulator, the RDI, is the place to check.

The three duties

1. The duty of care. You demonstrably manage cyber risks with appropriate measures: a current risk analysis, security of your network and information systems, a working incident procedure, business continuity and supply chain security. Demonstrably is the key word. Policy that sits in a binder but is not followed does not count as managing the risk.

2. The reporting duty. A significant incident triggers a tight schedule: an early warning within 24 hours, a full report with an initial assessment of severity and impact within 72 hours and a final report within a month. You will only meet those deadlines if you spot incidents quickly and have set up the reporting route in advance.

3. The registration duty. Organisations covered by the law register through mijn.ncsc.nl. Voluntary registration has been possible since October 2024; from 15 August it is mandatory.

The board is on the hook

The Cyberbeveiligingswet places responsibility squarely with the board. Directors must approve the security measures themselves, oversee their implementation and undergo training so they can assess the risks. For serious negligence, board members can be held personally accountable. Formally, the days of delegating cybersecurity to IT and looking the other way are over.

The fines underline the point: up to €10 million or 2 per cent of global annual turnover for essential entities, and up to €7 million or 1.4 per cent for important entities.

Where AI use meets the law

NIS2 never mentions AI, and yet AI sits right in the middle of it. AI tools are part of your attack surface and your data processing, and every AI vendor is part of the supply chain your duty of care must cover. An employee pasting customer data into a public chatbot creates exactly the kind of risk this law expects you to see, weigh and manage. And when it goes wrong, a data breach through an AI tool can be a reportable incident, with the 24-hour clock starting immediately.

The hard part: most organisations do not know which AI tools are in use. You cannot manage what you cannot see. How NIS2 and AI use come together is covered in NIS2 and AI.

What to do now

  1. Establish whether you are covered, and whether you count as essential or important.
  2. Register through mijn.ncsc.nl. The first formal step, and the easiest one.
  3. Update your risk analysis, including the supply chain and AI use in everyday work.
  4. Set up the reporting route. Who assesses an incident and who reports it? Can you meet the 24-hour deadline?
  5. Involve the board. Approving the measures and undergoing training are not formalities but legal requirements.

Getting a view of AI use does not have to be manual work. BeeSensible shows which AI tools are in use, how often and at what risk, checked against a catalogue of over 1,000 assessed tools.

AI register
AllowedUnder reviewNot allowedAll tools
Search by tool, vendor, domain or former name…
1,012 tools
DeepSeek
Critical · 90/100
Not allowed
Otter.aiNew
High · 68/100
No decision yet
Grok
High · 67/100
Not allowed
ChatGPT
High · 62/100
Allowed
Perplexity
Medium · 55/100
No decision yet
HeyGen
Medium · 53/100
No decision yet

Every tool, scored for risk

Over 1,000 AI tools, each scored from Low to Critical. Nothing is approved or blocked until someone decides.

More about AI Governance

Substantiating your measures without the busywork

The common thread in the Cyberbeveiligingswet is demonstrability. Taking measures is not enough; you have to be able to show they work. That is where most organisations get stuck, because collecting evidence is manual work that has to happen on top of the day job.

BeeSensible draws that evidence from the work itself: which AI tools are in use, which sensitive data was highlighted before it left the building and how it was handled. Always as counts, never the text and never an individual. The Accountability screen sets those figures, per month or quarter, against NIS2 Article 21(2)(a), (d) and (g) (Article 21(3) of the Cyberbeveiligingswet in the Netherlands), and shows for each provision what BeeSensible contributes and what lies outside its reach.

The first step takes an afternoon: register on mijn.ncsc.nl and write down which AI tools are in use today. The rest of the duty of care builds on that.

See

Which AI tools are actually in use, how often and at what risk. Every tool is scored against a catalogue of over 1,000 tools.

See the AI Governance module
Decide

Every tool's status is the organisation's own call: allowed, limited use, not allowed. The risk score is advice, never a verdict.

See the AI register
Help

That decision shows up while someone works: a notice on the tool, a highlight while they type. Blocking is not something BeeSensible does by itself.

See Realtime Privacy
Substantiate

The same observations become your evidence: six figures per period, next to the provisions of the GDPR, NIS2 and the AI Act. Never what was typed, never an individual.

See Substantiate
Dashboard · Accountability GDPRNIS2AI Act
  • Visibility of AI use

    Which AI tools were visited, and how many of them have a decision

    AI tools seen · Tools with a decision
  • Protection of sensitive data

    Critical detections, and how many were masked or removed before sending

    Critical detections · Adjusted
  • Supporting staff

    AI notices shown, and how often the alternative was opened afterwards

    AI notices shown · Alternative opened
The same observations, as six figures next to the provisions of the GDPR, NIS2 and the AI Act, showing for each provision what falls outside BeeSensible's scope. BeeSensible supports compliance; it does not make you compliant by itself.

Further reading: what NIS2 means for your AI use and what happens when you block AI without an alternative.

FAQ

Common questions

When does the Cyberbeveiligingswet take effect?

On 15 August 2026. The Dutch Senate adopted the law on 7 July 2026, together with the law on the resilience of critical entities. From that date, the duty of care, the reporting duty and the registration duty apply.

Does my organisation fall under the law?

The law covers more than 8,000 organisations across eighteen sectors, including energy, healthcare, transport, digital infrastructure and government. Rule of thumb: mid-sized and large organisations in those sectors are covered, from roughly fifty employees or ten million euros in turnover. Some organisations are covered regardless of size.

What does the duty of care involve?

You must demonstrably manage cyber risks with appropriate measures: risk analysis, security of network and information systems, incident handling, business continuity and supply chain security. Demonstrably is the key word: policy on paper is not enough.

How does the reporting duty work?

For a significant incident you give an early warning within 24 hours, a full report with an initial assessment within 72 hours and a final report within a month. Reports go to the CSIRT and the regulator.

How high are the fines?

Up to €10 million or 2 per cent of global annual turnover for essential entities, and up to €7 million or 1.4 per cent for important entities. On top of that, boards can be held personally accountable for serious negligence.

What does AI use have to do with NIS2?

AI tools are part of your attack surface, your data processing and your supply chain, three things the duty of care expects you to manage. A data breach through an AI tool can become a reportable incident, and AI tools nobody approved are a supplier risk you cannot manage as long as you cannot see them.

Portrait of Rens Timmermans from BeeSensible

Prefer someone to walk you through it?

Rens, BeeSensible

I'm happy to give you a twenty-minute tour. Leave your details and I'll find a time that suits you. You're also welcome to call or email.

Prefer to talk to someone first?

Want a live demo, or a quote for 100+ users? Leave your details and we'll be in touch within one working day.