AI in the workplace: Co-determination for union representatives

Artificial intelligence can be introduced as separate tools, but also as functions in systems the company already uses. For example, it can be Copilot in Microsoft 365, AI support in case management systems or functions that suggest answers, summarize meetings, analyse data or prioritise tasks.  

Some changes are barely visible at first. Gradually, however, they may have an impact on work tasks, responsibilities, competence needs, staffing and the organisation of work. 

The Government and the main organisations in the labour market have entered into a separate AI agreement (regjeringen.no) which states that the introduction of AI shall take place within the Norwegian working life model and through established arrangements for cooperation and co-determination.

The employer is responsible - union representatives have a say in the decision-making process

It is the employer's responsibility to ensure that AI is introduced and used in a responsible manner. The employer must, among other things, assess risk, safeguard privacy and information security, ensure necessary training and set a framework for its use. 

As a union representative, your task is to help ensure that the consequences are assessed before important decisions are made. Ask questions, ask for information and highlight the members' experiences and concerns before decisions are made. 

Ask for information early 

When AI can influence work tasks, working methods, organisation or the employees' work situation, you should ask for information and clarify the need for discussion. This may apply, for example, to the use of AI in case processing, documentation, customer service or planning of work.

This is relevant both when the company acquires a new tool, and when an AI function is added to systems you already use. The same applies if the employer plans to change work processes, organisation or guidelines as a result of AI. 

Start by asking the employer to explain why the solution is to be introduced. Is it about reducing time spent, managing larger workloads, increasing quality or freeing up time for other tasks? Which employees and tasks will be affected? 

This is in line with the AI agreement between the government and the social partners, which states that employees and union representatives must be involved when AI is introduced and put into use.

What does the introduction mean for the employees? 

AI can change both the tasks, the pace and the responsibilities of the work. Before new technology is used, you should therefore consider what the change means for your employees. The four questions below help you look at the consequences for people, not just the functions of the tool.

1) How are work tasks and organization affected? 

Some tasks may be simpler or faster, such as writing reports, gathering information or processing standard inquiries. Other tasks may change character or disappear. The introduction may also lead to new tasks related to control, quality assurance and follow-up of the solution. 

Therefore, ask the employer to describe how the work tasks are expected to change. Who will be affected? Will employees be given new roles or new responsibilities? Will expectations regarding pace, production or availability change? 

It is also important to clarify how professional judgement is to be used. The fact that an AI solution proposes an answer or an assessment does not mean that the proposal is always correct. Employees must still be able to assess information critically and take responsibility for decisions where necessary. 

2) Do employees have the right skills and support? 

New ways of working often create a need for new skills. When introducing AI, the employer and employee representatives should discuss what skills needs may arise in the longer term, and how these are to be followed up. 

Clarify who needs training in what before the solution is used. The needs will often be different for those who will use the solution on a daily basis, and those who will check results or follow up on quality, privacy and information security. Employees must know what the tool can be used for, where the boundaries are and how the results are to be quality assured.  

It should also be clear who employees can contact when something goes wrong, and how you will follow up on experiences, errors and the need for change. 

3) How is privacy and control safeguarded? 

Ask for an explanation of what data the solution uses and where it comes from. Is data from documents, emails, calendars, customer information, or other sources of information used? How is this data processed?  

You do not need to familiarize yourself with all the technical details. The important thing is to clarify how privacy and information security are assessed, who has access to what information, and what restrictions apply to the use of the solution. 

It should be clear what types of information can be entered into which AI tools, and which information should never be shared. Also clarify who has access to the data. 

4) Should AI be used for control and follow-up of employees? 

Some AI solutions can analyze work patterns, measure activity, or provide recommendations that can be used to follow up on employees.  

If a solution may have an impact on the assessment of individuals, there is reason to be extra vigilant. The employer should be able to explain what data is used, what assessments the system makes and how errors or biases can be detected.  

Ask the employer to explain what personal data the system has access to and what restrictions apply to its use. For example, there may be restrictions on how information is used for control, monitoring or assessment of employees. 

Also ask for information about how the assessments are documented, how the data quality is followed up and how human supervision should work. The person supervising the solution must have the necessary expertise and authority. 

It must also be clear who is responsible for the decisions that are made. AI can be a support tool, but it cannot take over the employer's responsibility for sound assessments and decisions.  

Good guidelines lead to safer use 

Clear guidelines make it easier to use AI in a safe and predictable way. Employees should know which tools have been approved, what information can be shared with the solutions, what tasks AI can be used for, and which tasks still require human assessment. 

The guidelines should also describe how errors and deviations should be handled, who is responsible for follow-up, and how employees can get help if they are unsure about their use. 

As a union representative, you should ask to be involved when such guidelines are to be prepared or updated. This increases the chance that the rules will be understood and actually applied. 

Address benefits and efficiency improvements early 

AI can make certain tasks faster and change how the work is distributed. If this reduces the time spent on documentation, reporting or other routine tasks, the company will often expect benefits.  

Ask the employer to explain what benefits they envisage, and how these may affect tasks, workload, organisation or staffing. Although the consequences may be uncertain early in the process, it is important to raise the topic before expectations are translated into plans and goals. 

In the agreement, both provisions on restructuring and efficiency gains may be relevant in the discussion with the employer. Check the fact box for what is specifically stated in your collective agreement. 

The KIC Regulation in brief

The AI Regulation is the EU's regulatory framework for artificial intelligence. It is based on a risk-based approach: The greater the risk an AI solution may have for safety, health or fundamental rights, the stricter the requirements apply.  AI used in recruitment or management of employees may be examples of use that must be considered separately.  

The AI Regulation has not yet been implemented in Norway. The Government is working to incorporate it into the EEA Agreement and implement it through a Norwegian AI Act. In any case, the Regulation will not replace the rules on the working environment, privacy and co-determination, but supplement these. 

Further help 

• Want to understand more about what AI is and how it affects the workplace? Take the digital course AI in the workplace – theory for responsible working life 
• Read the national guide AI assistants in working life – a practical guide (regjeringen.no)