If an algorithm classifies a low-risk customer as high-risk, or fails to detect a suspicious money laundering pattern, who is responsible? The system, the employee, or the management? More importantly, can the institution explain how the system reached its decision if requested by the regulatory authority?
These questions are no longer hypothetical. The financial world is transitioning from the adoption of artificial intelligence to the governance of artificial intelligence, where the ability to interpret automated decisions and understand the logic of smart models has become a fundamental requirement for risk management and compliance.
In Jordan, this shift coincides with a clear national trend towards enhancing the use of artificial intelligence within the Jordanian Artificial Intelligence Strategy 2023-2027, which aims to build a regulatory and technical environment that supports the adoption of these technologies in priority sectors, including the financial sector.
The Central Bank of Jordan has taken a significant step by issuing a regulatory framework for the use of artificial intelligence in the banking sector, establishing clear principles related to governance, risk management, transparency, data protection, and responsibilities associated with the use of these technologies.
Therefore, the compliance officer of the future will not only need to understand laws, instructions, and financial indicators but will also be required to understand the nature of the smart systems used by their institution, their limitations, their working mechanism, the data sources they rely on, and their susceptibility to bias or operational errors.
The real challenge does not lie in possessing artificial intelligence but in the ability to control and manage it responsibly. An algorithm whose decisions cannot be interpreted may turn from a tool for enhancing compliance into a new source of risk. An institution that does not understand the working mechanism of its smart systems may find itself unable to justify its decisions to regulatory authorities, customers, or even its senior management.
Financial institutions have successfully established a culture of "Know Your Customer" over the past decades, and it may be time to add a new principle to the modern compliance lexicon: "Know Your Algorithm". Trust in the digital future will not be built on the power of technology alone but on the strength of the governance that governs it.
In light of the accelerating digital transformation in Jordan, the question in the future may not be: "Does your institution use artificial intelligence?" but rather: "Does your institution know how the artificial intelligence it uses thinks?"
