Artificial intelligence may not be used for misleading marketing or to pressure customers

The Central Bank confirmed that it is not permissible for banks or insurance companies to use artificial intelligence to market various products in a misleading manner or to put pressure on customers, stressing the right of the consumer to request a human review and provide clear mechanisms for complaints, according to details published by the Central Bank, yesterday, on its website as part of a memorandum of directive guidelines to protect the consumer and ensure the responsible use of artificial intelligence in the financial sector.

He added that the application of artificial intelligence in the financial sector should not lead to discriminatory results, calling on banks to operate artificial intelligence systems under effective human supervision, especially in influential decisions.

The Central Bank defined a “high-impact decision” as any decision made using artificial intelligence that materially affects a customer’s access to a financial product or service, such as approving a loan or an insurance claim.

He continued: “The Board of Directors and senior management bear full responsibility for artificial intelligence systems and their results, including selecting models, developing them, disseminating them, being accountable for them, providing appropriate human resources, supervision, monitoring and ongoing management of them. Periodic reports must also be prepared and requested from the relevant management to senior management and boards of directors that address performance and risks.”

He said: “Boards of directors and senior management must ensure that risk committees and oversight functions (such as compliance, internal audit, and risk management) understand the processes driven by artificial intelligence, and are able to review and question the results when appropriate, with the adoption, deployment, and use of artificial intelligence considered an integral part of the risk management framework. A record (inventory) must be maintained of all artificial intelligence models, systems, or technologies that are developed or deployed, with compliance and training measures included within all functions responsible for their use, management, and monitoring, in line with standards. Forms Management and Forms Management Guidelines issued by the Central Bank (2022)”.

He added: “Artificial intelligence systems must be subjected to periodic testing (once a year or every time the model is upgraded, substantially modified, or a new model is introduced) to detect and address any unintended built-in biases or discriminatory results. The deployment of AI must reflect the organization’s ethical standards and code of conduct. Decisions made or supported by artificial intelligence must also be consistent with the duty to act honestly, fairly, and in the best interests of consumers.”

The Central Bank stressed that the level of human intervention must be proportional to the level of potential risks to the consumer, and the consumer has the right to request a human review or interpretation of decisions issued by artificial intelligence, and alternative arrangements must be provided when refusing to submit to an automated decision. Clear and easy channels for complaints must be maintained, and they must be addressed efficiently, confidentially, and in the shortest reasonable time. Artificial intelligence must not be used to target customers with inappropriate products, pressure sales practices, or misleading marketing.

The Central Bank stressed that when relying on external service providers or cloud service providers, due diligence must be conducted regarding governance, security, and data protection, and ensure the inclusion of contractual provisions that allow audit rights, access to information, and compliance with the requirements of the Central Bank. The justification for choosing the provider should be documented, annual cybersecurity reviews should be conducted by independent parties, a record of all models should be maintained, including those hosted by third parties, and excessive reliance on a single provider should be avoided if possible.

In a statement issued yesterday, the Central Bank said that it issued a memorandum of directive guidance on consumer protection and ensuring the responsible use of artificial intelligence and machine learning technologies by licensed financial institutions operating in the country, in a step that reflects its proactive supervisory approach and keeping pace with the rapid developments in digital transformation and smart financial services.

The memorandum aims to establish a clear framework that guides financial institutions towards employing these technologies in a safe and responsible manner. The guidelines define a set of basic principles that constitute a reference for the optimal use of these technologies, and include governance and accountability, fairness and reducing bias, transparency and avoiding ambiguity, and effective human supervision, in addition to data protection and privacy requirements.

The memorandum also includes provisions related to consumer disclosures, ensuring the clarity of the information provided in Arabic and English, and ensuring the availability of appropriate options for objection or withdrawal in high-impact cases resulting from automated decisions.

The Governor of the Central Bank, Khaled Mohammed Balama, said: “The memorandum of guidelines aims to establish a clear framework for the responsible use of artificial intelligence and machine learning technologies in the financial sector, in a way that enhances consumer protection and establishes the principles of governance and transparency, while emphasizing the importance of human supervision and data protection requirements.”

Expert: “Excessive customization” in artificial intelligence is not in the customer’s interest

Asim Jalal, a consultant in management sciences and information technology at G&K, said, “Artificial intelligence plays a major role in the banking sector, especially what is technically called (hyper-customization), meaning that the bank can provide detailed services that suit each individual customer exclusively according to his needs. For example, if the customer wants to obtain a loan, he can use (hyper-customization) to find out the appropriate loan for him, according to the information and data available about him, but on the other hand, that information can also be mobilized in some way to put pressure on the customer, or To market in a way that is in the interest of the bank and not in the interest of the customer.

Jalal added: “This feature allows banks to create a suitable model exclusively for each customer, even if the number of customers is in the millions,” pointing out that “excessive customization” depends on the customer’s data, and therefore the marketing team in the banks can use this data and information to understand the customer’s weak points, know his needs, and thus pressure him to accept products that he may not need or do not suit him.

• The Central Bank called on banks to operate artificial intelligence systems under effective human supervision, especially in influential decisions.