Legal Limits on the Use of Artificial Intelligence in Employer Decisions: The Role of Algorithmic Decisions in Labor Law

Technological advancements are bringing about profound changes in the workplace, just as they are in every other field, and are giving rise to transformations that must be examined from a labor law perspective. Artificial intelligence is being integrated into many systems, particularly in human resources processes, as it is believed to offer significant benefits in accessing qualified labor. As a result, AI-powered algorithmic management systems are playing an active role in employers’ decisions across a wide range of areas in the modern workplace—from hiring to termination, and from performance evaluations to salary determination. Consequently, this transformation has sparked legal debates regarding the applicability of classical and traditional labor law rules to the new order in employer-employee relationships.

Such AI applications, which are integrated into employers’ decision-making processes, raise new risks in many areas, particularly regarding the processing of employees’ personal data, job security, the principle of equality, and the privacy of private life. Therefore, the legal boundaries regarding the use of artificial intelligence in the workplace must be re-examined by taking into account both national and international regulations, and legal gaps must be addressed by comparing them with international standards. From the employer’s perspective, in this transformation, legal risks must be properly analyzed; if AI systems are to be integrated into any decision-making mechanism, the employer plays a crucial role in ensuring this is done through a preventive approach and in compliance with the law.

Applications of Artificial Intelligence in the Workplace

One of the most common areas where artificial intelligence is used in employer-employee relationships is the hiring process. Today, artificial intelligence can be utilized at every stage of the hiring process, and it is even possible for AI to make the final decision regarding whether a candidate is hired or not. Particularly in corporate settings, where companies receive thousands of job applications, ensuring the hiring of the most suitable candidate represents a significant workload. Employers are increasingly turning to AI support in this area to reduce both time and costs, lighten the workload, and reach a more diverse pool of applicants. 

Artificial intelligence is also utilized in the termination of employment contracts, particularly through the collection of performance-based data to enable algorithms to conduct performance evaluations; it is observed that AI plays a prominent role in identifying cases of low performance or absenteeism. One of the most common examples in performance tracking involves measuring the duration of calls made by call center employees, monitoring their tone of voice and time spent at the computer, and using this data for productivity assessments. 

In addition to these, artificial intelligence is also being utilized in areas such as work hours and shift scheduling, leave processes, monitoring workflow on the production line, occupational health and safety, and compensation and bonuses.

Legal Risks Posed by the Use of Artificial Intelligence in Employer Decisions

As mentioned above, the use of artificial intelligence in employer decisions entails certain legal risks. Violation of anti-discrimination laws constitutes one of these risks. While it is generally assumed that algorithms do not operate based on intuition, impressions, or biases, they may inadvertently reproduce biases present in the datasets used to train them. In other words, since algorithmic systems typically apply the same formula equally to all employees, they create the impression of appearing neutral. However, this apparent equality can be deceptive if the data feeding the system contains historical biases, as these biases may be reproduced.

However, the fact that the inner workings of algorithmic AI decision-making systems cannot be clearly and precisely understood from the outside also poses risks regarding transparency. This situation leaves both employers and employees in a state of uncertainty. If the employer cannot access the factors underlying the algorithmic system’s decision, situations may arise where the employer cannot sufficiently explain why an employee received a low score or was terminated. This raises the risk of conflicting with the principle of justified and verifiable termination expected under traditional labor law. 

On the other hand, the recording of conversations with employees—whether audio or video—for analysis, as well as the necessity to process and share various data, also poses risks regarding the processing of personal data. 

Furthermore, the use of AI systems in other areas such as occupational health and safety may also bring various additional risks to the fore. The fact that AI systems are not infallible, the ethical issues they may raise when used in areas such as employee monitoring, and the potential for legal liability arising from decisions supported by AI that could be deemed incorrect or one-sided are also legal risk topics that require detailed scrutiny.

International and National Legal Framework

European Union:

The European Union stands out as one of the legal systems that has come closest to establishing a comprehensive regulatory framework in the field of artificial intelligence. The EU Artificial Intelligence Act (AI Act), adopted in 2024 and gradually coming into effect, classifies artificial intelligence systems related to employment as “high-risk.”

Within this scope, algorithmic systems used in processes such as hiring, candidate screening, performance evaluation, task allocation, and termination of employment contracts are subject to specific obligations. These obligations include: ensuring systems are subject to a risk management framework, ensuring the quality of the data used, creating technical documentation, and fulfilling certain transparency requirements.

Furthermore, ensuring human oversight over the outputs of such systems is considered a fundamental requirement. In this context, decisions should not be left entirely to automated systems, and human intervention must be made possible when necessary.

On the other hand, the use of systems designed to detect employees’ emotional states, particularly in the workplace context, has been significantly restricted; certain biometric categorization practices have been banned.

In the assessment and policy documents published by the European Parliament’s Committee on Employment and Social Affairs in 2025, the risks of leaving critical employment-related decisions entirely to automated systems are highlighted, and it is recommended that regulations be developed to enable workers to request meaningful explanations regarding algorithmic decisions that affect them.

United States:

In the United States, it appears that a comprehensive and uniform federal-level regulation at the intersection of artificial intelligence and labor law has not yet been established. However, it is noteworthy that various regulatory initiatives have emerged at the state and local levels.

In this context, regulations adopted in the state of Colorado in 2024 impose obligations to prevent the use of artificial intelligence systems in a manner that could lead to discrimination. Similarly, in the state of Illinois, regulations requiring employers to inform employees when using artificial intelligence tools in hiring decisions have come into effect. In New York City, a requirement for independent bias review of automated decision-making tools is being implemented.

Judicial decisions have also begun to play a decisive role in the development of this field. In particular, in certain cases regarding the legal liability of AI providers, it has been indicated that these providers could be considered an actor acting on behalf of the employer under specific conditions. Court rulings have also begun to play a decisive role in the development of this field. In particular, certain cases regarding the legal liability of artificial intelligence providers have indicated that these providers may be considered actors acting on behalf of the employer under specific conditions. This approach raises an important area of debate regarding the possibility that liability may not be limited solely to the employer.

Assessment from the Perspective of Turkish Law:

It can be stated that there is currently no specific regulation in Turkish law focused on labor law that is unique to artificial intelligence. However, it can be argued that existing legislation contains various principles that indirectly limit algorithmic management practices, even if not directly. 

Under the Labor Code No. 4857, the employer’s right to manage is balanced by the obligation to protect the employee’s personal rights. In this context, it appears debatable whether decisions made solely based on algorithmic system outputs—particularly regarding the termination of an employment contract—meet the “valid cause” criterion stipulated by the Law in every specific case.

Under the Personal Data Protection Law No. 6698, algorithmic management practices are particularly significant in terms of the scope of data processing activities. Practices such as monitoring employee behavior, measuring performance, or processing biometric data necessitate a separate assessment within the framework of the principles of data minimization and proportionality. 

It can be stated that decisions in Turkish case law directly focusing on algorithmic management practices are still very limited. However, established case law regarding labor law emphasizes that performance evaluations must be based on objective, measurable, and verifiable criteria. In this context, it is likely that the question of whether the outputs of automated systems alone can serve as grounds for termination will become clearer in the coming period through judicial decisions. 

Preventive Measures in the Use of Artificial Intelligence Decision-Making Mechanisms in the Workplace

Undoubtedly, like any new technology, algorithmic artificial intelligence decision-making systems bring not only numerous advantages but also certain risks and challenges for legal systems. Employers must act by leveraging the undeniable benefits provided by artificial intelligence while addressing potential pitfalls such as discrimination, lack of transparency, privacy concerns, and over-reliance, and ensuring compliance with legal regulations.

If automated hiring processes using artificial intelligence systems are to be implemented, the employer must ensure that the algorithms and training datasets are free from personal biases and must verify that the system does not exhibit any distortion resulting from past instances of discrimination. In this context, technical documentation should be requested from the AI system provider regarding how the system was trained, which datasets it operates on, and whether bias tests are conducted regularly; furthermore, strategic provisions addressing risks arising from discriminatory outcomes should be included in contracts between the parties.

In addition, based on the European Union’s Artificial Intelligence Act, ensuring that decisions, outcomes, or recommendations generated by an AI system are subject to human oversight is a crucial step toward mitigating legal risks. Ensuring that such artificial intelligence systems are designed to be reviewable by human beings and that this aspect is carefully managed from the very beginning of the development process is a critical element. After all, protection against discrimination is a measure that must be addressed as early as the programming phase. Any solution developed to ensure that systems are designed from the outset to identify and prevent discrimination will be a beneficial step toward the future. It is also important, in this regard, that everyone with decision-making authority be informed about how the algorithmic systems they use work, what their limitations are, and under what conditions they should reject automatic recommendations, and that they receive training to gain a legal perspective.

However, the data that can be collected through artificial intelligence systems poses numerous risks regarding employee privacy and the protection of personal data. When using AI-based applications, it would not be inaccurate to state that there is always a risk of collecting and using more data than is necessary for the task at hand, depending on the analysis method employed. Furthermore, in hiring processes, the audio or video recording of interviews required for the use of AI systems may also infringe upon the data subject’s personal rights. For this reason, ensuring the AI-based applications used in hiring and business processes are used in compliance with data protection regulations, justifying data collection processes, fulfilling the duty to inform, and ensuring that only necessary data is processed is a highly important yet challenging process. In this regard, it is crucial that AI decision-making systems are programmed not to ask unauthorized or unnecessary questions during the establishment of an employment relationship, and that necessary data is carefully selected.

On the other hand, monitoring for purposes such as productivity and performance management is quite common among employers, and it is observed that this may include electronic communications, computer content, and social media activities.

Employers must act in compliance with data protection legislation in such data processing situations as well; they should take care not to use more data than necessary for legitimate purposes and not to create a conflict of interest, ensuring that the data is relevant, limited, and proportionate to the purpose for which it is processed, and that explicit consent and approval are obtained from the employee. 

It is also recommended that employers closely monitor the use of artificial intelligence and maintain comprehensive documentation. Keeping records of which data and parameters the system uses, when and on what grounds a decision was made, how human oversight was involved in that decision, and what actions were taken if any issues were detected in the system will be of great importance in any future legal proceedings. 

In light of all these considerations, employers should take care to use these algorithmic AI decision-making systems—which they will integrate into their decision-making processes—as a complementary tool rather than one that replaces human judgment; these systems must be subject to continuous oversight, updated in accordance with legal regulations, regularly audited by independent organizations, and used in compliance with principles such as ethics, equality, and privacy. 

From the employees’ perspective, it is also important to address the preventive measures that can be taken in this area. Under the KVKK, every employee has the right to know which of their personal data is being processed, by whom and for what purpose this data is being used, and how this processing affects their personal rights. Therefore, being aware of these rights, it is crucial to familiarize oneself in advance with the questions that can be asked and the objection mechanisms available in practice. 

Furthermore, just as it is crucial for employers to maintain documentation, it is equally important for employees to do so. Employees should retain every algorithmic decision made regarding them in their own records and preserve documents such as performance feedback emails or evaluation reports, as these are vital for any potential legal proceedings.

A low performance rating, employee selection based on an algorithmic filter, or an automatically initiated disciplinary process may be situations that require professional legal evaluation. With this in mind, employees can prevent potential harm by seeking advice from a specialized attorney, obtaining support through the legal department of their union (if they are members), and objecting to the manner in which their personal data is processed by contacting the data controller under the KVKK.

Conclusion 

In conclusion, artificial intelligence and algorithmic management systems are no longer a matter of choice in the business world but have become an almost de facto necessity for sustainable competition. The widespread adoption of artificial intelligence in the workplace and the expected increase in productivity are viewed by many experts as harbingers of a fundamental transformation. Consequently, under current conditions, the debate is no longer about whether to use artificial intelligence, but rather how and within what boundaries it should be used.

Employers remain responsible for the outcomes resulting from algorithmic tools and are obligated to ensure that the systems they use are fair and compliant; for employees, however, the issue is no longer merely a matter of individual rights but has become a systemic risk that simultaneously affects an increasingly broader segment of the workforce. Therefore, the need for both parties to proceed on a solid legal foundation is clear, and a preventive approach forms the basis of this foundation. 

In the near term, it has become an urgent necessity to address the legal gaps in this area through regulations centered on transparency obligations, human oversight standards, and clear appeal mechanisms for algorithmic decisions. At this juncture, legal regulations are of great importance not to halt the algorithm, but to ensure its fairness.  

Before employers integrate artificial intelligence systems into their operations, and when employees are subject to algorithmic decisions, seeking situation-specific professional legal support is critical to protecting against potential future risks, as each concrete situation involves unique conditions.   

Attorney Selin Ünverdi