Welcome to RegAcademy! Contact us for information on custom training and discount group rates.

RegAcademy

Please log in to access your purchased courses.

AI Risks and Opportunities for Financial Institutions (Hong Kong)

Learn how large language models work, why they hallucinate, and how financial institutions can use AI effectively. Explore key concepts including context windows, retrieval augmented generation, and agentic AI. Understand Hong Kong’s regulatory expectations for AI governance and cybersecurity, and discover practical approaches to building AI workflows with appropriate human oversight, security controls, and cost management.

Suggested Credit: 1 Hour

Topics: AI, Governance, Cybersecurity

How Large Language Models Work

  • Neural Networks and Model Training
  • Model Scale and Emerging Capabilities
  • Hallucinations and the Risks of Trusting AI Outputs

Key AI Concepts and Capabilities

  • Tokens and Context Windows
  • Retrieval Augmented Generation (RAG)
  • Tools and Model Context Protocol (MCP)
  • Agentic AI

Hong Kong Regulatory Expectations

  • Senior Management Accountability and AI Governance
  • Model Risk Management and Validation
  • Controls for High-Risk Uses of AI
  • Cybersecurity, Data Protection, and Third-Party Risk
  • AI-Enabled Cyberattacks and Incident Response

Building AI Solutions for Financial Services

  • Defining Tasks, Inputs, and Outputs
  • Structured Workflows and Agentic Harnesses
  • Human Oversight, Guardrails, and Access Controls
  • Testing, Auditability, and Continuous Improvement
  • Model Selection, Cost Management, and Prompt Quality
Featured
Start the Course

Let’s get started

How Large Language Models Work Part 1 – How Large Language Models Work Key AI Concepts and Capabilities Part 2 – Hong Kong Regulatory Expectations and Cybersecurity Hong Kong Regulatory Expectations and Cybersecurity Part 3 – Hong Kong Regulatory Expectations and Cybersecurity Building AI Solutions for Financial Services Part 4 – Building AI Solutions for Financial Services