Standard Chartered is integrating artificial intelligence (AI) into its processes, paying particular attention to privacy regulations.

Privacy by Design

The bank considers data privacy as a fundamental element in the design and deployment of AI systems. David Hardoon, Global Head of AI Enablement at Standard Chartered, emphasizes how privacy requirements define the type of data that can be used, the level of transparency required, and how systems are monitored.

Deployment Challenges

The transition from pilot projects to real operating environments presents significant challenges. Data quality becomes a critical issue when AI systems draw on various upstream platforms, each with its own structure and potential inconsistencies. Privacy rules impose further constraints, such as the use of anonymized data for model training, which can affect performance.

Data Sovereignty and Localization

Geography plays a crucial role in the deployment of AI systems. Data protection laws vary from region to region, and some countries impose strict rules on data localization. Standard Chartered adopts a flexible approach, with both centralized and local deployments, depending on the regulations in force. Data sovereignty is a key consideration, especially for systems that process personal information.

Human Oversight and Standardization

Transparency and explainability of AI systems have become essential. Human oversight remains fundamental, especially when automated decisions affect customers or regulatory obligations. Standard Chartered focuses on standardization, creating pre-approved templates and data classifications to simplify the application of privacy and governance rules.

Staff training is another crucial aspect. Employees must be aware of what data can be used, how it should be handled, and what limits must be respected.