Since the publication of ChatGPT at the latest, the topic of Artificial Intelligence (AI) has been on everyone’s lips. The chatbot has shown to a broad public what AI can do and that AI will find its way into people’s everyday lives.
The current high level of attention paid to AI provides a reason to examine the significance of AI in terms of labor law. The following article is dedicated to this topic. The authors see four fields of action that HR managers should be aware of.
Labor law assessment of AI applications
AI is currently used in companies mainly as a supplement and support in the form of individual AI applications. From the perspective of labor law, the question of the legal admissibility of individual AI application always arises. In this regard, protecting the personal data of applicants and employees, employment data protection, anti-discrimination law and co-determination in companies must be taken into account.
Active sourcing using AI
Active sourcing refers to measures designed to identify suitable candidates for a position. Using AI, HR managers can have suitable applicants suggested to them from databases for the profile they are looking for. This involves accessing data on potential candidates from professional networks such as LinkedIn or Xing. As long as the data processing is limited to data from professional networks that have been made public by the candidates themselves, this data processing is generally permissible.
Robo recruiting
Things become more difficult when the recruiting process is largely taken over by AI. It would be technically possible, for example, to record job interviews or telephone interviews with applicants. AI can evaluate the applicants’ language and then make statements about candidates’ character traits. While such tools may be widespread in the USA, they are difficult to reconcile with German data protection law. This is true even if the applicant has consented to the collection and analysis of this data.
Workforce planning using AI
AI tools can also optimize staff deployment. As far as digital scheduling is concerned, scheduling employees (e.g., through duty rosters) is familiar territory. AI can be used to optimize the scheduling of employees based on knowledge previously gained about them.
The evaluation of employee performance data by AI, on the other hand, is a source of concern for employee representatives. However, these concerns are not particularly justified. Warnings and dismissals of employees for poor performance are recognized in German labor law. However, they hardly play any role in practice because poor performance can barely be proven in court. The use of AI will make it more difficult for low performers in the future. This is because AI can analyze qualitative and quantitative benchmarks on the basis of which personnel decisions can be made in a comprehensible manner.
AI-controlled instructions
Through the use of AI, instructions to employees are increasingly no longer issued directly by human superiors, but by AI-supported applications. There is nothing intrinsically wrong with this from a labor law perspective.
However, if the AI evaluates the work and performance behavior of employees and subsequently issues instructions to optimize the work process, the monitoring obligations of the “natural” supervisor will increase. Since instructions must always comply with equitable discretion, this limit must be observed. Employers cannot rely on the fact that AI tools issue instructions to employees on their own.
Selection decisions using AI
AI tools are also used to terminate employment. For example, AI-supported analyses can be used to make the social selection to be made when downsizing and to determine the employees who will be laid off. However, employers must be able to explain the criteria on which the social selection is based for subsequent dismissal protection proceedings. Simply referring to the selection process by AI is not sufficient.
In particular, Article 22 (1) of the General Data Protection Regulation (GDPR) stands in the way here. According to this, employees have the right not to be subject to a decision based solely on automated processing – including profiling – which produces legal effects concerning them or similarly significantly affecting them.
Impact of AI deployment on HRM
It is not possible today to foresee in detail how AI will change the world of work. What is certain, however, is that AI will bring about comprehensive change processes.
The automation of tasks through digitalization and the use of AI will be at the center of this. New activities will emerge through AI, new forms of work will develop. Jobs that can be easily replaced by AI applications will be lost. The implementation of this adaptation process is probably the greatest challenge that AI will pose to HR managers.
German labor law provides suitable instruments for the emerging steps. For example, the recruitment of skilled workers must be prepared, employment contracts must be adapted, transfers and dismissals must be carried out. The legal design of qualification agreements will also become more important than in the past if employees are to be made fit for the new AI working world.
Legislator reactions to AI
The use of AI is leading to increasing reactions from legislators.
Artificial Intelligence Act
For example, the EU Commission has proposed the Artificial Intelligence Act (AIA), a regulation establishing harmonized rules for artificial intelligence (COM/2021/206 final). The draft provides for a risk assessment for AI systems, primarily from the point of view of the protection of personal rights. Depending on the risk classification, certain applications will be prohibited. AI applications used in the employment relationship, including its justification, are classified as high-risk AI systems. This means that although these applications are not prohibited in principle, they will be associated with special obligations for the employer.
Article 22 (1) GDPR
As has already been mentioned, Article 22 (1) of the GDPR is particularly important. It must be expected that the idea behind this will become a central principle for AI in the employment relationship. Decisions taken by means of AI vis-à-vis employees will have to be traceable – at least indirectly – to a natural person or a body.
Section 80 (3) BetrVG
If there is a works council at a company, the introduction and use of an AI system is usually subject to the works council’s co-determination pursuant to section 87 (1) 6 of the German Works Constitution Act (BetrVG). This is because AI tools are technical devices designed to be able to monitor the behavior or performance of employees. It makes sense for the parties involved to regulate the relevant aspects in a company agreement.
One of the few provisions of labor law that already explicitly mentions AI is section 80 (3) of the BetrVG. Only companies with a works council are affected. If the works council has to assess the introduction or application of artificial intelligence in order to carry out its tasks, the works council may call in an external expert – at the employer’s expense.
Using legal tech to handle labor law cases
A fourth area of action where the use of AI has an impact concerns the way HR managers, employment law departments and employment lawyers work. Lawyers should not fool themselves. Their work can also be replaced – in some areas – by AI. Simple legal questions can already be answered using legal tech. AI tools are suitable for getting an initial overview at least. Increasingly, legal value judgements in the run-up to personnel decisions are being prepared by AI. AI can also draft opinions in labor court proceedings, as far as standard cases are concerned.
As a result, AI will change the cooperation between human resources experts and labor lawyers. However, it is not expected that AI will replace individual labor law advice tailored to the individual case – especially in complex labor law cases and special matters, which frequently occur in labor law.
Conclusion
Individual AI applications should be subjected to an admissibility check under labor law before they are implemented. Not everything that is technically possible is permitted under German law.
Probably the greatest challenge posed by the use of AI is to implement the expected transformation process in the world of work. AI is going to permanently change the world of work. Labor law provides suitable instruments for this.
The density of legal regulations for AI is going to increase. It is important to carefully observe which regulations the legislator creates at national and European level. It is expected that decision-making powers in companies may only be outsourced to AI to a limited extent.
Finally, AI is going to have an impact on HR work itself, the work in labor law departments and the cooperation between HR and labor lawyers. In this respect, too, AI will take over work steps.
michael.riedel@advant-beiten.com
