AI e abusi di mercato: le leggi della robotica si applicano alle operazioni finanziarie?: E‘ il tema affrontato nel nuovo Quaderno Giuridico della Consob (30 maggio 2023)

The Consob Legal Journal No. 29, published in May 2023, focuses on the theme of AI and market abuses in the financial context.
It discusses the application of robotics laws to financial operations in the context of AI and market abuses and highlights the challenges posed by the use of advanced AI algorithms and their impact on legal categories. The Journal explores the question of whether AI systems should have separate legal personality and examines the complexities of attributing responsibility and liability to AI algorithms. It also addresses specific legal issues in different areas of law, such as corporate governance and antitrust, and emphasizes the challenges of regulating the use of AI in capital markets while balancing market integrity and investor protection. Overall, it acknowledges the difficulties in addressing these complex legal problems and highlights the need for careful consideration of AI’s implications in various legal contexts.
The widespread adoption of AI systems by businesses poses challenges to existing laws and regulations as certain decisions, including those with detrimental effects, can now be made by both humans and AI systems. As a result, there is a growing demand for specific regulations on AI from national institutions and the European Union. The European Commission has proposed a regulation on AI (please see EventID#11011 for details), that aims to promote the development of AI applications while also distinguishing AI systems based on the risk they pose to human rights. The Commission combines various risk protection techniques, including the precautionary principle for systems with unacceptable risk and the prevention principle for systems with high risk. In the context of the financial market, the application of AI systems has transformed certain services such as high-frequency trading, robo-advisory, and credit scoring.
Furthermore, the journal explores the distinction between weak and strong AI systems in the financial market.
Weak AI systems rely on predefined instructions provided by manufacturers, programmers, or users, while strong AI systems have the ability to autonomously learn and generate independent and unpredictable outputs compared to the initial inputs.
The widespread adoption of these technologies in the financial sector presents new challenges for legal protection. The current regulatory framework primarily focuses on human behavior, both active and passive, without fully addressing the financial misconduct related to market abuses facilitated by AI systems.
In particular, strong AI systems require the adoption of new criteria for assigning responsibility, which raises questions about the application of the principle of technological neutrality in financial regulation.
The document proposes three possible alternative solutions to address the harmful behaviors of strong AI systems, but each solution presents specific challenges that are further discussed in the Journal.
These include attributing legal personality to advanced AI systems, establishing objective liability for those who put AI systems into operation, and considering the socialization of the costs of harm caused by AI systems. The EU’s proposed Regulation on Artificial Intelligence, which combines the precautionary and preventive principles based on risk, could provide a solution that balances AI utilization with fundamental rights protection. Implementing responsibility for producers, programmers, or users should be based on the creation of a risk rather than their knowledge of the event. However, any chosen solution should strike a balance between technological development and ensuring adequate market functioning and protection of legal rights affected by AI agents.