PAKISTAN ZINDABAD

SBP Nears Completion of AI Guidelines for Banks

KARACHI: The State Bank of Pakistan (SBP) is finalizing comprehensive guidelines to ensure the responsible and ethical use of artificial intelligence (AI) in the financial sector.

These upcoming guidelines are designed to foster trust, transparency, and accountability in AI-driven services, while safeguarding consumer rights and interests.

According to the SBP’s Financial Stability Review 2024, AI adoption is accelerating across industries worldwide, including the banking sector, where financial institutions are increasingly integrating AI technologies into their operations.

In Pakistan, banks are following suit, primarily using AI for robotics, process automation (handling structured tasks), virtual assistants (enhancing customer service), and machine learning (for fraud detection and risk management).

AI Adoption Survey

In 2024, the SBP conducted a survey of 55 regulated entities—including conventional and Islamic commercial banks, microfinance institutions, digital banks, Electronic Money Institutions (EMIs), and Payment Service Operators/Providers (PSOs/PSPs)—to assess the state of AI adoption, the challenges involved, and future outlook.

The survey found that around 50% of these institutions have either implemented or are in the process of developing AI solutions. These tools are being used across a wide range of functions, including fraud detection, customer service, marketing, credit risk evaluation, and process automation.

Addressing Environmental and Operational Risks

The report highlights that, alongside their benefits, AI systems come with risks that need careful management. One emerging concern is the environmental impact of AI, particularly its carbon footprint. SBP emphasizes the need to treat this as a specific risk category within banks’ overall risk management frameworks.

“The first step is acknowledging the environmental risks that come with AI,” the report states.

Given that many AI technologies used in banking are energy-intensive due to real-time data processing and high accuracy demands, banks are encouraged to measure the emissions of AI models throughout their life cycles and adopt energy-efficient alternatives.

International standards like IFRS S1 and S2 now require companies to disclose information on sustainability-related risks, including climate exposures. In Pakistan, the adoption of these standards will be rolled out gradually, starting with listed companies that meet specific thresholds for assets, revenue, or workforce size.

The SBP also recommends that banks establish clear limits for acceptable environmental impact from AI, allowing them to balance performance and sustainability.

Systemic Risk Considerations

Another critical focus is the systemic risk AI may pose to financial institutions and their customers. Increasing reliance on technology, along with a limited number of AI suppliers, could heighten vulnerability to operational disruptions, cyberattacks, or supply chain issues.

Additionally, widespread use of similar AI models across institutions may lead to greater market correlations, potentially amplifying asset price volatility.