Eyes on productivity: AI-driven employee monitoring

Swiss labor law generally prohibits surveillance systems that monitor employees’ behavior. But there are some exceptions to this rule. For instance, such systems can be justified for reasons such as safety, quality or performance optimization, provided that they do not unreasonably infringe upon employees’ health or personality rights.

In practice, distinguishing between permitted and prohibited monitoring can be challenging. Therefore, implementing a legally sound surveillance system calls for a thorough case-by-case analysis. Labor law and relevant court rulings outline three primary conditions for an acceptable surveillance system:

1. The employer must have a clear and predominant interest, such as safety or performance optimization;
2. Monitoring must strike a balance between the employer’s operational needs and employees’ reasonable privacy expectations; and
3. Employees should be involved in planning, establishing operating times and determining data retention duration.

In line with the principle of proportionality (see point (2) above), video surveillance, due to its more intrusive nature, is considered justified only under narrow circumstances, for example in areas with increased security needs (e.g., near hazardous machinery). Similarly, highly automated surveillance systems evaluating employee data are permitted only under strict conditions and depending on their specific purpose.

Hence, while AI systems could be used for certain performance optimizations, detailed analysis of employees’ productivity or, more generally, their activities in the workplace on a continuous, periodic or sampling basis for example through automatic vision or motion data analysis would generally be prohibited under labor law.

Surveillance systems typically involve processing employees’ personal data and must therefore adhere to Swiss data protection principles. Amongst others, employers must inform employees in advance of any collection of personal data, duly meet their rights if exercised and respect Privacy by Design principles when architecting AI-driven surveillance systems.

Swiss labor law imposes additional limits on processing employee data. Generally, personal data may only be processed for the performance of the employment contract. Further processing may require specific justification, such as the employee’s consent. However, relying on employee consent can create challenges for employers. The hierarchical structure within workplaces may lead to situations where employees feel compelled to agree, potentially undermining the legal validity of their consent.

In addition, there are heightened considerations when it comes to systematic and comprehensive evaluation of personal aspects or the large-scale processing of sensitive employee data (e.g., facial, fingerprint or iris scans to record employees’ time and attendance). In these cases, employers should be particularly mindful of additional legal limitations. This becomes especially pertinent when an AI-driven surveillance system goes beyond providing operational insights and is relied upon to make decisions directly affecting an individual employee (e.g., terminations). Employers must ensure that such decisions are made with careful consideration and human oversight, to prevent potential infringements on the employees’ rights and ensure compliance with legal requirements.

Non-compliance with these legal requirements can have serious consequences, including:

1. If evidence for an employee’s dismissal is obtained based on unlawful surveillance systems, such evidence may be inadmissible in a dispute relating to the termination of an employment relationship.
2. Recording acoustic and visual backgrounds without consent of the employees concerned may constitute a criminal offense.
3. Violation of certain provisions of Swiss data protection law (e.g., duty to provide information) can result in fines of up to CHF 250,000 for the responsible individuals.

Implementing a legally sound surveillance system requires a thorough understanding of both legal and technical requirements on a case-by-case basis. This is especially vital when it comes to the strict requirements surrounding video and highly automated systems, potentially utilizing AI.

Establishing clear policies around functionality, type and timing of surveillance as well as involving employees in the planning phase is highly recommended. In addition, it seems essential to closely follow and consider the ongoing regulatory developments (e.g., the EU AI Act) when designing and implementing AI-driven surveillance systems.