In Hong Kong, financial institutions are increasingly adopting generative artificial intelligence (genAI), aiming for enhanced productivity and operational efficiency.

A study commissioned by the Hong Kong Institute for Monetary and Financial Research (HKIMR), the research arm of the Hong Kong Academy of Finance (AoF), surveyed 55 entities from the banking, insurance, and wealth and asset management sectors between October 2024 and January 2025.

The findings revealed that 75% of respondents had either already implemented at least one genAI use case, were piloting and designing use cases, or were exploring potential investment areas. This ratio is expected to increase to 87% within the next three to five years.

These figures mark a significant increase from 2023. According to Finastra’s annual global survey on the exploration and roll out of new models and technologies, 38% of Hong Kong organization had started rolling out genAI at the time. This suggests that the adoption of genAI is progressing steadily across the city’s financial services industry.

Larger institutions lead adoption

The HKIMR study also found that larger institutions are adopting genAI at a faster pace. Among surveyed firms, 83% of large firms had rolled out at least one genAI use case or were taking steps towards adoption, compared to 63% of small firms.

The study also found that in 2024, the primary implementations of genAI in financial services in Hong Kong were focused on internal and non-customer-facing applications, with virtual assistants for employees being the most common genAI use cases. Customer-facing applications, meanwhile, were restricted to customer chatbots and enhanced sales and marketing.

This trend aligns with how genAI is perceived by the industry. 75% of the surveyed financial institutions reported viewing genAI as a tool to enhance productivity and operational efficiency, followed by 53% who viewed genAI as empowerment for innovation and decision-making.

These findings are consistent with a previous 2023 study by the Hong Kong Monetary Authority (HKMA), KPMG China and Quinlan & Associates, which found that 80% of genAI use cases in the financial services sector focused on internal applications aimed at automating tasks, improving workflows and sharing knowledge.

This suggests that financial institutions may still be taking a cautious, internal-first approach to genAI, testing and refining the technology in less risky environment before deploying it externally.

Risks and challenges hindering adoption

Despite growing implementation of genAI, several risks and challenges are still hindering adoption. When adopting genAI, financial institutions cited model accuracy and performance (95%), model transparency and explainability (65%), and data privacy and security (64%) as the top three risk management considerations.

This underscores the importance of accuracy and data integrity and security in ensuring the safety and functionality of genAI. This is especially critical in an industry like financial services were errors can carry significant financial, regulatory, and reputational costs.

Financial institutions were also mindful of new methods of cyber attacks that leverage genAI. When asked about the three most critical cybersecurity risk that genAI creates or exacerbates, 75% and 60% of survey respondents ranked deepfake and artificial identities (75%) and phishing attacks (60%) as the top risks.

Deepfake scams on the rise

Deepfakes, which use AI to create highly realistic digital forgeries, have proliferated and grown in sophistication over the past years amid increased accessibility and affordability. According to data from Signicat’s The Battle Against AI-Driven Identity Fraud report, deepfake fraud attempts have grown by 2,137% in the financial services industry in the last three years.

Just last month, a finance director of a multinational corporation nearly lost over US$499,000 through a business-related impersonation scam involving digital manipulation, the Straits Times reported.

The scammer posed as the company’s CFO and arranged a fake video conference involving deepfake representations of the company’s executives. The victim was then instructed to transfer over US$499,000 from the company’s HSBC bank account to a money mule account under the control of the scammer.

The money was eventually recovered by the Anti-Scam Centre (ASC) of the Singapore Police Force and the Hong Kong Police Force’s Anti-Deception Coordination Centre (ADCC).

In January 2025, Hong Kong police arrested 31 people of a local syndicate that produced deepfake romance and investment scams, the South China Morning Post (SCMP) reported. The ring had defrauded victims in Taiwan, Singapore and Malaysia, with criminal proceeds exceeding HK$34 million (US$4.37 million).

Featured image credit: edited from freepik