LONDON (May 30): Banks and investment firms in the EU cannot shirk their board responsibilities and legal obligations to protect customers when using artificial intelligence (AI), the European Union's securities regulator said on Wednesday, in its first statement on AI.
The European Securities and Markets Authority (ESMA) on Thursday outlined how financial institutions regulated in the 27 countries can use AI in their day-to-day operations without breaching the EU's MiFID securities law.
ESMA said that while AI holds promise for improving investment strategies and customer service, it also carries inherent risks that could have significant implications for the protection of retail investors.
“Importantly, company decisions, whether taken by humans or by AI-based tools, remain the responsibility of the governing body,” ESMA said.
“Central to the use of AI in investment services is an unwavering commitment to acting in the best interests of clients, an overarching requirement that applies regardless of the tools firms decide to adopt in delivering their services.”
ESMA said the statement covers not only cases where banks or investment firms have developed or adopted AI tools themselves, but also the use of third-party AI technologies such as ChatGPT and Google Bard, with or without the direct knowledge or approval of senior management.
“Company management needs to have a proper understanding of how AI technologies are applied and used within their companies and ensure appropriate oversight of these technologies,” ESMA said.
The statement focuses on MiFID compliance and is separate from the EU's landmark rules on AI that come into force next month and set a potential global benchmark for the technology used in business and everyday life.
Efforts are also underway at a global level by the Group of Seven (G7) countries to establish guardrails to ensure the safe development of rapidly evolving technologies.
