Cos warns of the risk that artificial intelligence amplifies banking crises

In a room at the Willard Hotel in Washington, on the block next to the White House, in a place steeped in history and with somewhat old decoration, the governor of the Bank of Spain, Pablo Hernández de Cos, dedicated himself this Wednesday to talk about technology, robots, machine learning and artificial intelligence. At the conference organized by the Institute of International Finance (IIF), the association of large banks chaired by Ana Botín, Cos has warned that artificial intelligence will force a rethinking of financial supervision and that, if not controlled, it can amplify the banking crises.

The governor of the Bank of Spain spoke with the cap of president of the Basel Committee on Banking Supervision. Cos has stressed that digital innovation will further boost cross-border and cross-sector financial interconnections. “Safeguarding global financial stability will require continued collaboration to ensure an appropriate gold standard regulatory and supervisory approach to oversee the use of AI/ML in the banking and other sectors,” the governor said.

Artificial intelligence and

(AI/ML) in the financial sector is neither good, bad nor neutral, has been his thesis, following one of Melvin Kranzberg's laws of technology. He has exposed the positive and negative aspects and has warned of the risks and opportunities it poses, but underlining that although it is early to draw conclusions, it will have a relevant impact.

“My main message is that the use of AI in banking poses significant prudential and financial stability challenges. We have yet to see how AI/ML performs over a full financial cycle, and this could take some time. If left unchecked, these patterns could amplify future banking crises. But these challenges and limitations are not insurmountable, as long as central banks and supervisory authorities adapt to this new reality and collaborate effectively,” he indicated.

Such collaboration is likely to extend to a wide range of authorities, beyond central banks and banking supervisors, given the continued growth of non-bank players and the blurring of regulatory boundaries, Cos said. He has also noted that some advocate a “technoprudential” approach to AI governance, to oversee system-wide financial risks.

Digitalization report

“When it comes to banking, it is essential that banks anticipate and monitor the risks and challenges posed by artificial intelligence and machine learning at both the microeconomic and macroeconomic levels and incorporate them into their daily risk management and governance mechanisms” , the governor has insisted.

Cos, however, has emphasized not getting carried away by technology and robotics, but rather maintaining control. Paraphrasing Melvin Kranzberg's final law on technology, he said: “Despite attempts to draw a robotic future for banking, the truth is that human beings—and particularly human judgment—can and should continue to play a role. integral role.”

In his opinion, central banks and supervisory authorities must also anticipate and mitigate possible risks and vulnerabilities of the banking system. Cos has announced that the Basel Committee on Banking Supervision will soon publish a more comprehensive report on the digitalization of finance and the implications for regulation and supervision.