The development of new technologies and artificial intelligence is certainly streamlining or even automating an increasing number of processes. Indeed, many solutions are being widely used, including in HR departments. Unfortunately, some of them raise significant legal concerns, and regulations are still being drafted. What changes should employers be preparing for and what should they be addressing right now?
The world needs rules for artificial intelligence
The US and China are already working on AI laws, but it is the European Union that is closest to passing a relevant piece of legislation.
This is because in Europe, in addition to heated discussions, concrete work is being done on an EU regulation on artificial intelligence.
It is important to note that this will be a regulation, which means that once its comes into force, it will be directly applicable in all Member States, without the need for national legislation.
And, by the way, the General Data Protection Regulation (GDPR), which did not require implementation in the form of national laws, works on a similar basis.
Human Resources: a high-risk area
The regulation defines conditions for the use of AI systems in the field of employment in the broad sense (situations such as recruitment, promotion, dismissal decisions, etc.), which has been classified as a high-risk area.
This classification is based on the fact that the use of these systems has a direct impact on the lives and finances of most people.
The European Commission also notes that such systems may perpetuate historical patterns of discrimination based on gender, age, origin, etc., which is another reason for considering them as high-risk systems. Potential violations of data protection and privacy rights are also a significant risk.
Classifying employment as a high-risk area will entail additional obligations for those using AI systems.
EU regulation on AI in HR
The draft regulation requires that in high-risk areas, an appropriate system for managing these risks must be put in place and be regularly updated.
Employers will be required to take appropriate measures to manage these risks in accordance with the numerous requirements under the regulation.
In addition, systems in place should be tested in order to implement the most appropriate risk management measures. Such testing should be carried out before the system is placed on the market, and also ‘as appropriate’.
The regulation also requires that systems are transparent and enable users to interpret the output, and that they are subject to effective human oversight.
Most of the provisions of the proposed regulation only set out the objective to be achieved, leaving the appropriate measures to define the assumptions arising from the law up to the providers and users of the artificial intelligence.
The draft regulation itself suggests that employers wishing to support their processes with AI will need to prepare for a number of organisational steps, if only because the regulation provides for penalties of up to EUR 30,000,000 or, if the infringement is committed by a company, up to 6 % of its total worldwide annual turnover for the preceding financial year – whichever is higher.
Amendment to the Trade Union Act
Polish lawmakers are currently working on an amendment to the Trade Union Act of 23 May 1991 with plans to extend the list of information that employers must provide at the request of trade unions. As of September 2022, the draft is at the stage of first reading in committees.
The proposed amendments would require employers to provide information on the parameters, rules and instructions on which algorithms or artificial intelligence systems relevant to decision-making are based and which may affect conditions of work and pay, access to and retention in employment, including profiling.
With regard to this additional right of trade unions, there have been concerns that the extent of the information to which trade unions would have access could breach business confidentiality.
This is a significant risk due to the potential for disclosure of company know-how. Therefore, in order to balance the interests of both employers and trade unions, a non-disclosure obligation should be introduced together with extended access to information.
Artificial intelligence in the field of HR: current status
Does the fact that legislation directly addressing AI has yet to take effect mean that currently employers have complete discretion in this area? Absolutely not!
After all, it is important to remember that any action taken by an employer, and therefore any use of tools, including those based on artificial intelligence, must not lead to unequal treatment or be based on discriminatory criteria.
It is therefore worth reviewing them, especially from this point of view.
How to do this? How can your organisation adapt to the planned legal changes?
Contact us. We will advise you on the best steps to take in any situation.
