AI and the Works Council: 10 Questions to Ask in 2026

AI is no longer an experiment in many organisations. Employees use ChatGPT, Copilot and other AI tools, HR systems are gaining intelligent features and more and more business software uses algorithms to analyse information or support decisions.

That also makes AI a topic for the works council.

Its impact goes far beyond choosing a new piece of software. AI can affect jobs, workload, privacy, employee assessment, recruitment and the way work is organised.

A large part of the European AI Act also became applicable on 2 August 2026. This makes it a good time for works councils to put AI permanently on the agenda.

But where do you start?

These ten questions can help the works council have a practical conversation with management.

1. What AI is our organisation actually using?

Do not start with legislation. Start with an inventory.

AI is increasingly embedded in software that organisations already use. Examples include Microsoft 365, HR systems, recruitment tools, customer-service software, planning tools and analytics platforms.

Instead of only asking:

"Which AI systems have we purchased?"

also ask:

• Which existing systems contain AI features?
• Which AI tools are employees using themselves?
• Which systems process employee data?
• Which systems make recommendations or support automated decisions?
• Which new applications are currently planned?

Without an overview, it is almost impossible for the works council to properly assess the consequences.

2. Why do we want to use AI?

"Working more efficiently" is not a complete business case.

Ask which specific problem the organisation is trying to solve.

Is the goal to save time? Reduce costs? Address staff shortages? Improve quality? Support decision-making?

Just as importantly, how will the organisation determine whether the implementation has actually been successful?

A clear objective makes it much easier to assess later whether the benefits outweigh the consequences for employees.

3. What will change for employees?

This may be the most important question for the works council.

The Dutch Social and Economic Council, the SER, identifies potential effects of AI on areas including jobs, job content, autonomy, skills, working conditions and personnel policy.

Questions could include:

• Will certain activities disappear?
• Will new tasks be introduced?
• Will employees' responsibilities change?
• Will employees have more or less autonomy?
• Which new skills will be required?
• Will training be provided?
• Will staffing levels be affected?

Do not only consider the situation immediately after implementation. Some consequences may only become apparent several months later.

4. Does the works council have a right to advise or consent?

Not every AI application is subject to the same procedure under the Dutch Works Councils Act, the WOR.

The works council has a right to advise on significant financial, economic and organisational decisions. For AI, this can become relevant when technology forms part of a significant reorganisation or change to business processes.

The right of consent can become relevant when AI forms part of an employee-related policy, for example concerning employee assessment, working conditions, the processing of employee data or systems used to observe or monitor employees.

An AI project can also contain several different elements. Advisory and consent rights may therefore both be relevant.

Do not wait until the system has already been purchased. Employee participation should be involved during the preparation of the decision.

5. Are employees being monitored or assessed by AI?

This deserves particular attention from the works council.

Examples include software that:

• measures productivity;
• analyses conversations;
• compares performance;
• tracks attendance;
• selects applicants;
• ranks employees;
• analyses behaviour.

Where arrangements involve the processing of employee personal data or systems capable of monitoring employees, the works council's right of consent may apply.

The GDPR and, depending on the application, the AI Act may also be relevant.

The technical question "can the system do this?" can therefore be just as important as "do we intend to use it this way?"

6. What data is being used?

AI systems depend on data.

The works council can ask:

• What information is collected?
• Does it contain personal data?
• Where is the data stored?
• How long is it retained?
• Who can access it?
• Is the data used to train AI models?
• Is information shared with external suppliers?
• Can employees access or correct their information?

Pay attention to information that can indirectly reveal something about an employee as well.

A system does not need to explicitly display an "employee score" to make conclusions about performance or behaviour possible.

7. Who checks the AI?

An AI system can provide a recommendation, but somebody must remain responsible for what happens with that recommendation.

Ask how human oversight is organised.

Can an employee challenge an AI-generated outcome? Can a manager deviate from a recommendation? Are errors recorded? Who investigates complaints?

Especially when AI affects people, it needs to be clear who is ultimately responsible for the decision.

"The system decided" should never be the end of the discussion.

8. Can we explain how decisions are made?

Employees should understand when AI influences a process that affects them.

This is particularly important in areas such as recruitment, performance assessment, scheduling or task allocation.

Ask whether the organisation can explain:

• what role AI plays;
• what information the system uses;
• how much influence the AI output has;
• who makes the final decision;
• how an employee can challenge an outcome.

New transparency obligations under the AI Act also apply to certain AI systems from 2 August 2026.

9. Are employees adequately prepared to use AI?

The AI Act has contained obligations concerning AI literacy since February 2025.

This does not mean every employee needs to become an AI specialist. Organisations do, however, need to take the knowledge and experience of people working with AI into account and ensure they sufficiently understand how the technology should be used and what risks it may create.

Useful questions for the works council include:

• What training will employees receive?
• Do employees know what information they may and may not enter into AI tools?
• Can they report mistakes or unsafe situations?
• Are managers trained to evaluate AI-generated output?
• Does the organisation have an internal AI policy?

AI literacy is not just about learning how to write good prompts. It is also about continuing to think critically.

10. How will we monitor AI after implementation?

Approval of a system should not mean closing the file.

AI software changes frequently. Models are updated, new functionality is introduced and suppliers can change the way their products work.

Agree on evaluation points in advance.

For example, after three, six and twelve months.

Review:

• results;
• errors;
• complaints;
• effects on workload;
• effects on jobs;
• privacy;
• employee usage;
• unexpected consequences.

Also agree when a significant change needs to be submitted to the works council again.

Put AI permanently on the agenda

Perhaps the most important step is to stop treating AI as a one-off IT project.

The SER recommends discussing upcoming AI policy with the works council at an early stage, including during the consultation under Article 24 of the Dutch Works Councils Act.

This gives the works council an opportunity to participate before the most important choices have already been made.

For example, every quarter you could discuss:

1. Which new AI applications are being investigated?
2. Which pilots are currently running?
3. Which applications have been introduced?
4. What effects are employees experiencing?
5. Which upcoming decisions may require employee participation?

This moves the works council from responding to individual systems towards structural participation in the organisation's digital strategy.

A practical AI checklist for the works council

Has your works council received an AI proposal? Start with these questions:

• What does the system do?
• Why do we need it?
• Which employees will be affected?
• Which activities will change?
• What data will be processed?
• Are employees monitored, selected or assessed?
• Who checks the output?
• How can employees challenge a decision?
• What training will employees receive?
• How will the system be evaluated after implementation?

A works council does not need to understand how an AI model is technically constructed.

It does need to understand what the technology means for the people expected to work with it.

Using AI within the works council itself

AI can also support the work of the works council.

It can help summarise large files, prepare questions, analyse employee feedback or create a first draft of meeting minutes.

The same principles remain important: confidentiality, control over data and human review of the results.

Ordamo was developed around this way of working. Meetings, documents, topics, advice and consent requests and employee feedback can be managed in one central environment, with AI functionality built specifically around employee participation processes.

AI does the groundwork. The works council remains in control.

Sources

This article uses information from sources including the Dutch Social and Economic Council (SER), the Dutch government and the European Commission concerning employee participation, the Dutch Works Councils Act and the European AI Act.

This article provides general information and does not constitute legal advice.