Regulation of artificial intelligence
Pro Research Analysisby 
Searched over 200M research papers
Defining Artificial Intelligence for Regulation
A key challenge in regulating artificial intelligence (AI) is defining what exactly is being regulated. There are many definitions of AI, but the most enduring is Turing’s, which focuses on systems that display intelligent behavior by analyzing their environment and acting with some autonomy to achieve goals. However, regulators must be careful not to anthropomorphize AI, as this can lead to misunderstandings about its capabilities and risks Finocchiaro2023Harasimiuk2021.
Approaches to AI Regulation: Sectoral vs. General
There is ongoing debate about whether AI should be regulated as a whole or through sector-specific rules. The European Union (EU) has chosen a general approach with its AI Act, aiming to create a unified regulatory framework while also considering geopolitical factors. Other countries and regions are experimenting with both broad and targeted regulations, depending on the specific risks and applications involved Finocchiaro2023Minbaleev2018De Almeida2021+1 MORE.
The European Union’s AI Act and Enforcement
The EU’s AI Act, adopted in 2024, is a leading example of comprehensive AI regulation. It establishes the European AI Office to oversee enforcement and implementation. The Act requires market surveillance, human oversight, post-market monitoring, and mandatory reporting of serious incidents. High-risk AI systems face strict obligations before being placed on the market, and exceptions are clearly regulated Harasimiuk2021Minbaleev2018Reed2018.
Risk-Based and Ethics-Centered Regulation
Modern AI regulation often uses a risk-based approach, focusing on protecting fundamental rights and societal values such as fairness, privacy, and non-discrimination. The EU’s framework, for example, requires that high-risk AI systems meet strict standards to ensure they do not harm individuals or society. However, defining and operationalizing these values can be challenging due to their normative ambiguity, requiring ongoing political and scientific debate to ensure legitimacy and legal certainty Minbaleev2018Sridhar2019Cuéllar2022.
Regulatory Tools and Governance Models
Effective AI regulation may require a mix of tools, including licensing regimes, mandated transparency, formal verification of system behavior, and rapid intervention capabilities. Traditional regulatory models, such as delegating oversight to expert agencies, remain relevant but must be adapted to address the unique challenges of AI, such as the opacity of deep learning systems and the unpredictability of their behavior Kerekes2023Orwat2024.
Challenges and Problems in AI Regulation
Key problems in regulating AI include uncertainty about future developments, potential threats to security and human rights, and the need for legal guarantees to ensure safe functioning. Different countries are addressing these issues by focusing on both immediate concerns and long-term strategies for system-wide regulation. Ensuring transparency, accountability, and human control are recurring themes in legislative initiatives worldwide De Almeida2021Judge2024Cuéllar2022.
The Importance of Trust and Social Values
Building public trust in AI technologies is essential. Regulatory frameworks emphasize transparency, human oversight, and ethical standards to foster trust. The integration of ethical guidelines, such as those for trustworthy AI, is seen as a foundation for future laws and industry self-regulation Minbaleev2018Kerekes2023Reed2018.
Incremental and Adaptive Regulation
Some experts argue that it is too early for a comprehensive system of AI regulation and recommend working incrementally within existing legal frameworks. New regulations should be introduced only when current laws are inadequate, with a focus on transparency and accountability, especially for high-risk applications .
Conclusion
Regulating artificial intelligence is a complex and evolving challenge. Effective regulation requires clear definitions, a balance between general and sector-specific rules, risk-based approaches, and strong enforcement mechanisms. Ongoing dialogue, ethical considerations, and adaptive governance are crucial to ensure that AI technologies are safe, trustworthy, and aligned with societal values Finocchiaro2023Harasimiuk2021Minbaleev2018+7 MORE.
Sources and full results
Most relevant research papers on this topic
Artificial Intelligence Regulation: a framework for governance
The Artificial Intelligence Regulation framework provides a comprehensive tool for guiding policy decisions on AI regulation, aiming to achieve fairness, freedom, and long-term sustainability.
How should we regulate artificial intelligence?
AI regulation should be incremental and based on existing legal and regulatory schemes, with producers providing transparency to explain AI decisions.
DOI