Artificial Intelligence law at Serbia
1. Overview of AI Law in Serbia
Serbia currently does not have a dedicated, comprehensive law regulating AI. There is no AI-specific statute, no formal regulatory body solely for AI, and no reported case law directly addressing AI liability or compliance.
Instead, AI is currently regulated indirectly through:
Existing general legislation, such as:
Data protection laws
Consumer protection laws
Civil liability laws
Criminal law
Government strategies and ethical guidelines, which are policy instruments rather than legally binding rules.
The legal system is in a transitional phase, preparing for the introduction of AI-specific regulation in the near future.
2. Policy Framework and Ethical Guidelines
Serbia has adopted AI development strategies covering:
Promotion of AI research and innovation.
Responsible deployment in public and private sectors.
Alignment with ethical principles such as transparency, accountability, fairness, human rights, and non-discrimination.
Coordination via advisory bodies and councils for AI governance.
Ethical guidelines exist, focusing on principles like:
Human oversight of AI systems
Prevention of bias and discrimination
Data privacy and protection
Transparency in algorithmic decisions
However, these guidelines are not legally enforceable, so failure to comply does not yet trigger legal penalties.
3. Application of Existing Law to AI
Because there is no AI-specific statute, any legal issues arising from AI are handled under general legal rules:
Civil liability: Companies or individuals can be held responsible for damages caused by AI under general tort law. For example, if an AI system causes financial harm, the responsible human or entity may be liable.
Criminal liability: If AI is used to commit fraud, defamation, or other offenses, natural or legal persons involved can be prosecuted under general criminal law. There is no direct criminal liability for the AI itself.
Data protection and privacy: AI systems processing personal data must comply with existing data protection regulations. Violations can lead to fines or sanctions against the operator.
Consumer protection: AI tools used in commercial settings must comply with general consumer rights laws, including obligations related to product safety and transparency.
4. Case Law Status
Currently, Serbian courts have not issued decisions specifically related to AI. There is no precedent establishing:
Liability for decisions made solely by AI
Obligations for AI transparency
Specific regulations for AI system deployment
In practice, disputes involving AI would be assessed using existing laws on negligence, liability, or data protection. Courts have not yet developed a framework for AI-specific jurisprudence.
5. Current Challenges
The absence of AI-specific legislation in Serbia creates several challenges:
Legal uncertainty: Companies and individuals cannot be certain how AI-related actions will be treated by courts.
Regulatory gaps: Certain harms, such as algorithmic bias or automated discrimination, may not be adequately addressed under current laws.
Enforceability: Ethical guidelines are non-binding, meaning actors can ignore them without direct legal consequences.
Rapid AI adoption: The law risks lagging behind technological development, particularly in sectors like healthcare, finance, public administration, and justice.
6. Ongoing Developments
Serbia is actively preparing for formal AI regulation:
A working group has been established to draft the country’s first AI law.
The law is expected to define:
Categories of AI systems (e.g., high-risk AI)
Obligations for developers, providers, and users
Transparency and accountability requirements
Liability frameworks for AI-caused harms
Compliance with ethical standards
This law will likely be influenced by the EU AI Act but adapted to Serbian context.
Once enacted, courts will begin to interpret and apply the new AI rules, gradually producing AI-specific case law.
7. Summary Table
| Area | Current Legal Approach | Notes |
|---|---|---|
| AI Development & Deployment | Guided by strategy and ethical guidelines | Non-binding, advisory in nature |
| Civil Liability | General tort law applies | Responsible human or entity held liable |
| Criminal Liability | Existing criminal law | AI itself is not a criminal subject |
| Data Protection | General data protection law applies | AI operators must comply |
| Case Law | None specific to AI | Courts rely on general law for disputes |
8. Outlook
Serbia is in transition from soft law to binding AI regulation.
No AI-specific case law exists yet; disputes are resolved under general laws.
Future legislation will define responsibilities, rights, and liabilities for AI use, bringing legal clarity and potential for new jurisprudence.
Until then, individuals and companies must rely on existing civil, criminal, and data protection laws to navigate AI risks.
RELATED Blog
comments