Can the Use of Artificial Intelligence Be Regulated in Employment Contracts? The Legal Nature and Contractual Limits of the Employment Relationship

The increasing use of artificial intelligence tools in the workplace has necessitated a reassessment of the mutual rights and obligations of employers and employees. While employees utilize AI for tasks such as text writing, data analysis, software development, customer communication, reporting, translation, and design, employers face new risks regarding productivity, data security, the protection of trade secrets, and the accuracy of the resulting work.

Therefore, employers want to regulate in employment contracts which AI tools employees can use, whether company data can be uploaded to these systems, how AI outputs will be monitored, and the consequences of violating these rules.

The use of artificial intelligence can be regulated in employment contracts. However, adding an unlimited clause to the employment contract such as "The employer may monitor and modify the employee's use of artificial intelligence in any way it deems necessary" does not automatically render all practices legally compliant.

The validity of a contractual provision must be determined by considering together the employer's right to manage, the protection of the employee's personality, the protection of personal data, the prohibition of fundamental changes in working conditions, the obligation to work personally, the duty of loyalty, and the principle of proportionality.

Are there any specific regulations regarding the use of artificial intelligence in Turkish law?

As of July 31, 2026, there is no comprehensive artificial intelligence (AI) law in Türkiye that fully regulates all aspects of AI use in employment relationships. Draft laws aimed at ensuring the safe, transparent, fair, and accountable use of AI systems are currently before the Turkish Grand National Assembly (TBMM) committees. Therefore, AI applications in the workplace are evaluated based on the existing Labor Law, the Turkish Code of Obligations, the Law on the Protection of Personal Data, and other relevant legislation.

This does not mean that employers cannot regulate the use of artificial intelligence. Employers can establish rules regarding the use of technology within the scope of the organization and execution of work. However, these rules cannot contradict mandatory legal provisions and the fundamental rights of the employee.

Can the use of artificial intelligence be regulated in an employment contract?

Regulations regarding the use of artificial intelligence;

• Directly to the employment contract,
• A protocol that serves as an annex to the employment contract,
• Information security and artificial intelligence usage policy,
• To the personnel regulations,
• According to the job description,
• Privacy and intellectual property commitment

can be added.

In practice, instead of directly writing detailed technical rules into the employment contract, it would be more appropriate to define the general legal framework in the contract and regulate the details in an "Artificial Intelligence Usage Policy" that serves as an annex to the employment contract.

However, simply having an employee sign a general clause stating that they "must comply with company policies" should not be considered sufficient. The policy must be presented to the employee, its contents must be understandable, and prohibited behaviors must be clearly indicated.

Will the introduction of artificial intelligence change the nature of employment contracts?

The fundamental elements of an employment contract are the performance of work, wages, and a relationship of dependence. The fact that an employee performs their duties with the assistance of artificial intelligence does not, as a rule, eliminate these elements.

Even if it utilizes artificial intelligence;

• To work within the employer's organization,
• To comply with the employer's instructions,
• To perform the act of working,
• To receive payment in return

It continues.

Therefore, the use of artificial intelligence does not transform an employment contract into a contract for work, an agency contract, or any other type of contract. In most cases, artificial intelligence is a helpful tool used in performing the work, like a computer, calculator, or professional software.

However, while the use of artificial intelligence may not change the legal type of contract, it can affect the manner in which the employee's work obligation is fulfilled, the job description, the supervision system, and the responsibility.

The Worker's Obligation to Work and Artificial Intelligence

According to Article 395 of the Turkish Code of Obligations, unless otherwise agreed in the contract or due to the circumstances, the employee is obligated to perform the assigned work personally.

An employee's use of artificial intelligence does not, in all cases, constitute a breach of their personal work obligations. What is important is that the employee maintains their personal contribution, professional control, and responsibility for their work.

For example, an employee's interaction with artificial intelligence;

• Creating a text draft,
• Correcting spelling mistakes,
• Grouping data,
• Generating ideas,
• Getting code suggestions,
• Summarizing a long text

Using it for that purpose does not mean that the entire business has been transferred to a third party.

Conversely, an employee's act of having artificial intelligence prepare their assigned work entirely without any verification, presenting the AI ​​output as their own work, or leaving a task requiring professional evaluation to the system's outcome, could constitute a breach of their duty of diligence and care.

Especially in professions requiring personal professional evaluation, such as law, medicine, engineering, accounting, and human resources, it should be mandatory for artificial intelligence outputs to undergo human review.

Can employers mandate the use of artificial intelligence?

According to Article 399 of the Turkish Code of Obligations, the employer may make general regulations and give specific instructions regarding the performance of work and the conduct of employees in the workplace. Employees are obliged to comply with these regulations and instructions to the extent required by the principles of good faith.

In this context, the employer;

• The use of a specific enterprise artificial intelligence tool,
• The control of the produced outputs by humans,
• Company data should only be processed in approved systems
• Reporting on the use of artificial intelligence,
• Confidential information should not be uploaded to open systems

He/She may want to.

In many cases, updating existing software or providing employees with a new tool that makes their work easier may fall within the employer's management rights.

However, if making the use of artificial intelligence mandatory significantly alters the employee's duties and responsibilities, increases their workload, makes their work constantly monitored, or changes the wage and performance system, the matter can no longer be considered solely within the scope of management rights.

When does the application of artificial intelligence constitute a fundamental change?

According to Article 22 of the Labor Law No. 4857, an employer can only make substantial changes to working conditions established by the employment contract, personnel regulations, or workplace practices by notifying the employee in writing. Changes not accepted in writing by the employee within six working days are not binding on the employee.

Not every AI intervention represents a fundamental change. The assessment should consider whether the intervention worsened or aggravated the employee's previous situation.

The following practices may be considered a fundamental change depending on the specifics of the case:

• Significantly changing the employee's job description,
• An employee is removed from their previous area of ​​expertise and assigned to oversee artificial intelligence systems
• Significantly increased targets and workload due to artificial intelligence,
• Linking the bonus system to algorithmic performance scores,
• Continuous monitoring of an employee's screen, keyboard, voice, or behavior,
• Reducing an employee's salary based on a productivity score calculated by artificial intelligence,
• Placing all responsibility for AI errors on the employee,
• The employee is transferred to a significantly lower-skilled role.

Conversely, the use of an enterprise artificial intelligence system, provided free of charge and with training, to ensure the company's data security, may fall within the scope of management rights if it does not burden the job description and working conditions.

The Supreme Court's approach also holds that changes falling within the scope of management rights or based on valid reasons cannot be considered fundamental changes in working conditions; however, practices that worsen the worker's conditions are subject to Article 22 of the Labor Law.

Are the general and unlimited authority clauses in the employment contract valid?

Artificial intelligence provisions, prepared in advance by the employer and added to contracts for a large number of employees, may constitute general terms and conditions of business under Articles 20 and subsequent articles of the Turkish Code of Obligations.

For general terms and conditions to be considered part of a contract, the employee must be explicitly informed of their existence, given the opportunity to learn their content, and accept the terms. Provisions that are irrelevant to the nature of the contract and the specifics of the work may be deemed omitted.

Therefore, provisions such as the following are open to legal debate:

"The employer can monitor all of the employee's digital activities without any restrictions."

"The employer can directly implement any decisions made by artificial intelligence systems on the employee."

"The employee is responsible for all damages arising from the use of artificial intelligence, even if they are not at fault."

"The employer can unilaterally change the employee's duties, wages, and working methods according to the needs of artificial intelligence systems."

The employer's reservation of the right to amend the contract does not grant unlimited and unchecked power to make changes. This power must be exercised objectively, in accordance with the principle of good faith, and in a way that does not disproportionately harm the employee's rights.

Are the provisions regarding the use of artificial intelligence sufficient under the Personal Data Protection Law (KVKK)?

Adding a clause regarding the use of artificial intelligence to an employment contract does not automatically make the processing of employees' personal data lawful.

Through the employer's artificial intelligence system;

• The employee's correspondence,
• The commands he gave to the artificial intelligence,
• The documents he prepared,
• Working hours,
• Performance data,
• Audio or video recordings,
• Customer meetings

Collecting this information may constitute a personal data processing activity.

In this case, the employer must specify the purpose of data processing, the legal basis, the data categories, the persons to whom the data is transferred, and the retention periods; and inform the employees in accordance with Article 10 of the KVKK (Personal Data Protection Law). The employee also has the right to object to any unfavorable outcome resulting from analysis performed exclusively by automated systems.

The general clause in the employment contract stating "The employee consents to the processing of all their personal data by artificial intelligence systems" is insufficient.

Whether explicit consent was given freely should also be evaluated due to the economic and hierarchical dependency between the employee and the employer. The Personal Data Protection Board acknowledges that explicit consent may not be considered valid in cases such as listing employees who have not consented or associating negative consequences with not consent.

Can an employer monitor an employee's use of artificial intelligence?

Employers may monitor employees' use of artificial intelligence on corporate systems to a certain extent for legitimate purposes such as protecting trade secrets, ensuring information security, regulatory compliance, and oversight of the work.

However, the monitoring activity;

• It must be based on a specific and legitimate purpose,
• This should be explained to the employee in advance
• It should be relevant to the purpose and limited in scope
• It should be implemented after evaluating whether a less intrusive method is available
• One should not interfere with the essence of private life.

In the Personal Data Protection Board's decisions regarding corporate email monitoring, it is important that employees are informed in advance about the purpose, scope, and conditions of the monitoring; that the data processing is based on legitimate grounds such as legitimate interest or the establishment, exercise, and protection of a right; and that it is proportionate.

Therefore, the justification that "the employee is using the employer's system" does not allow for the unlimited monitoring of all screen activity, correspondence, and private content.

Trade Secrets and Client or Customer Information

One of the most significant risks in using artificial intelligence is the uploading of company data to third-party AI platforms.

By the employee;

• Customer lists,
• The contracts,
• The source codes,
• Financial statements,
• Health data,
• Personnel information,
• Client files,
• Commercial information that has not yet been made public

Uploading this to an open artificial intelligence system could lead to a violation of confidentiality obligations, personal data protection provisions, and obligations regarding the protection of trade secrets.

Employment contracts or AI policies should specifically indicate which categories of information cannot be uploaded to AI systems. Instead of using a general statement like "confidential information cannot be uploaded," providing examples of types of confidential information increases predictability for employees.

Who is responsible for the accuracy of AI output?

Artificial intelligence systems can produce inaccurate, incomplete, or false content. Therefore, employment contracts can include a clause obligating the employee to review the AI ​​output.

However, an employer cannot automatically hold an employee liable for all damages resulting from an AI system after forcing them to use it without providing adequate training.

When assessing responsibility;

• Whether the employer has chosen an appropriate and safe system,
• Whether the employee received training,
• Whether the terms of use are clear,
• Whether the employee checked the output,
• Whether the error was predictable or not,
• Whether the employee is at fault,
• The nature of the job and the employee's level of expertise

They should be evaluated together.

The employer's duty to protect and care for the employee also continues. According to Article 417 of the Turkish Code of Obligations, the employer is obligated to protect the employee's personality, show respect to the employee, and ensure an orderly environment in the workplace that conforms to the principles of honesty in the employment relationship.

Could a violation of the Artificial Intelligence Policy be grounds for termination?

An employee's violation of artificial intelligence usage rules may result in disciplinary action or termination, depending on the nature of the incident. However, not every violation constitutes a direct reason for justified termination.

For example, the employee;

• Uploading trade secrets to a publicly accessible artificial intelligence system,
• Sharing customer personal data without permission,
• Creating fake documents or records using artificial intelligence,
• Knowingly using incorrect outputs without checking them
• Persistently acting in defiance of the employer's clear and lawful instructions

It can lead to more serious consequences.

Conversely, direct termination may be considered disproportionate if the usage policy is unclear, the employee is not trained, similar violations go unpunished for other employees, or the violation does not cause any harm.

In evaluating termination, consideration should be given to whether the rule was clear and previously communicated, the employee's fault, the severity of the breach, the resulting damage, the employer's consistent practice, and the principle that termination should be a last resort.

What topics should be regulated in an employment contract?

A legally valid and enforceable artificial intelligence regulation should clarify at least the following points:

1. Permitted AI tools and accounts for use,
2. Which tasks are limited to its use?
3. Information that is prohibited from being uploaded to artificial intelligence systems,
4. Protection of personal data, trade secrets and customer information,
5. Human control of artificial intelligence outputs,
6. The notification process to be followed when encountering incorrect or unlawful outputs,
7. Under what conditions can records of artificial intelligence usage be monitored?
8. The purpose, scope, and retention period of the monitoring activity,
9. Whether or not the work prepared using artificial intelligence should be reported to the employer,
10. Intellectual property and usage rights,
11. Training to be provided to the employee,
12. Proportional sanctions that can be applied in case of rule violations,
13. The procedure for changing the policy,
14. The employee has the right to object to automated decisions and erroneous data.

The clause added to the contract should be clear, understandable, and appropriate to the employee's job. Instead of using the same general text for every employee, it is more appropriate to prepare differentiated policies based on job and risk level.

Assessment in terms of the European Union Artificial Intelligence Regulation

The European Union Artificial Intelligence Regulation classifies certain AI systems used in the recruitment, task assignment, employment relationship decision-making, monitoring, and performance evaluation of employees as high-risk systems. The regulation's approach emphasizes safeguards such as risk management, data quality, record keeping, transparency, and human oversight.

Although the regulation does not directly apply to every Turkish employer, it should be examined separately for companies operating in the European Union, employing workers in the EU, or providing services to the EU market. Furthermore, the principles of transparency and human control in the regulation constitute an important standard of compliance for employment contracts and company policies to be prepared in Türkiye.

Conclusion

The use of artificial intelligence may be regulated in an employment contract or in a policy and protocol supplementing the employment contract. Such a provision, as a rule, does not alter the legal nature of the employment contract. Even if the employee benefits from artificial intelligence, they continue to work within the employer's organization and receive wages.

However, incorporating artificial intelligence into an employment contract is not merely a technical arrangement. Depending on the scope of its use, the employee's job description, their direct work obligations, duties of care and loyalty, performance system, personal data, and method of supervision may all be affected.

The use of a new artificial intelligence tool within the normal course of business may fall within the scope of management rights. However, applications that significantly affect an employee's duties, salary, workload, performance metrics, or personal life may constitute a fundamental change in working conditions and may be subject to the written notification and written acceptance procedure stipulated in Article 22 of the Labor Law.

The artificial intelligence clause in the employment contract;

• You shouldn't grant unlimited monitoring privileges
• It should not be used in place of the legal basis for processing personal data
• It should not eliminate all responsibility for the employee or place all risk on the employee
• It should not grant the employer unlimited and unilateral power to make changes
• It should not eliminate the employee's right to human oversight and to object.

A legally sound approach is to regulate artificial intelligence not as a decision-making authority that replaces employees and obscures accountability, but as a work tool subject to clear rules, human oversight, data security, and proportionality.