Can we use Artificial Intelligence (AI) to select employees for redundancy?

Artificial intelligence (AI) is increasingly becoming part of our HR processes. Among other things, the technology can be used to analyse employee data, compare competencies and support decisions concerning organisational changes. In connection with major restructurings and workforce reductions, it may therefore be tempting to use AI to identify which employees the company will need going forward. But can we use AI to select employees for redundancy? And what legal considerations should we, as employers, be particularly aware of? The short answer is that, as a general rule, we may use AI as a decision-supporting tool, but we cannot simply leave the actual decision on redundancy to the technology.

Where do we draw the line when using AI?

When we, as employers, face a restructuring or workforce reduction, we often need to make difficult assessments of employees’ competencies, qualifications and future relevance to the business.

AI can be a useful tool in this process. For example, we may use the technology to organise information about employees’ competencies, compare qualifications against the company’s future needs or identify relevant factors that should be considered in the selection process.

However, specific restrictions apply if we wish to use AI to carry out the actual selection.

Under Article 22 of the General Data Protection Regulation (GDPR), employees generally have the right not to be subject to decisions based solely on automated processing, including profiling, which produce legal effects concerning them or similarly significantly affect them.

A decision to dismiss an employee will normally fall within the scope of this provision.

This means that, as a general rule, we cannot allow an AI system to independently determine which employees should be selected for redundancy. Although the GDPR provides for certain exceptions, these will rarely be applicable in the context of ordinary workforce reductions.

The key requirement is therefore that AI is used solely to support the decision-making process, and that the final selection is based on a genuine and independent human assessment.

In other words, simply approving the recommendations generated by an AI system is not sufficient. We must review the assessments, critically evaluate the results and, where necessary, override the system’s recommendations.

We must still be able to justify our dismissal decisions

When using AI as a decision-supporting tool, the general principles of employment law continue to apply.

We must still be able to demonstrate that the selection of employees for redundancy is based on objective and relevant criteria, and that the decision does not constitute unlawful discrimination.

Challenges may arise, for example, if an AI system ranks employees based on performance, length of service, absence or competencies without us having sufficient insight into how the assessment was carried out.

We must therefore pay particular attention to whether the data and criteria used by the AI system are relevant and accurate, and whether the system’s results may give rise to a risk of direct or indirect discrimination on grounds such as age, gender or disability.

A recommendation generated by an AI system does not in itself demonstrate that a dismissal is objectively justified.

What about the EU AI Act?

In addition to the GDPR, we must also be aware of the EU AI Act, which introduces specific requirements for the use of AI systems, including in the employment context.

AI systems intended to make or support decisions concerning employment relationships, including dismissals, will generally be classified as high-risk AI systems.

This entails a range of specific requirements relating to, among other things, risk management, documentation, transparency and human oversight.

Under the current timetable, the relevant requirements for high-risk AI systems in the employment context will apply from 2 December 2027.

 

If we already use, or are considering implementing, AI-based tools in our HR processes, we should therefore consider both the existing data protection rules and the forthcoming requirements under the AI Act.

What should we pay particular attention to?

If we wish to use AI in connection with a restructuring or workforce reduction, we should in particular ensure that:

  1. The selection criteria are objective and relevant.
    We must determine in advance which data and criteria the AI system may use and how these support the company’s future needs.
  2. A genuine human assessment is carried out.
    We must be able to review and, where necessary, override the AI system’s recommendations before making the final decision.
  3. We comply with the GDPR.
    Among other things, we must ensure that the processing of employee data has a lawful basis and that the requirements relating to transparency, data minimisation and, where applicable, data protection impact assessments are met.
  4. We prevent discrimination.
    We must consider whether the system’s underlying data or assessments could result in unlawful discrimination against particular employees or groups of employees.
  5. We document our decisions.
    We must be able to explain how AI was used, which assessments we carried out ourselves and why the employees concerned were selected for redundancy.

What is our recommendation?

AI can be an effective tool for creating an overview and supporting more structured decision-making processes in connection with restructurings and workforce reductions.

However, before using AI for such purposes, we should ensure that clear internal guidelines have been established and that there is a common understanding of which decisions AI may support and which assessments we must carry out ourselves.

AI can help us make decisions about redundancies – but ultimately it remains our responsibility as employers to justify those decisions and stand behind them.

Do you consider dismissals or workface reductions?

BAHR advises on all aspects of dismissals, restructurings and mass redundancies – from the initial considerations and employee selection process to the practical implementation. We also provide advice on the use of AI in these processes, helping you establish a sound basis for your decisions.

We are currently experiencing a high demand on restructurings and mass redundancies. Please do not hesitate to reach out for a non-binding chat, should you consider dismissals or a restructuring.

 

Share aticle to
Loading video ...
close