概要

This webinar gives professionals at licensed corporations and registered institutions a practical foundation in generative AI: How large language models (LLMs) actually work and where the risks sit. Using a factory analogy, we explore how to build defensible AI workflows that capture the productivity upside without crossing regulatory lines.

The topic is timely because AI adoption inside the industry is accelerating while the SFC, HKMA, and PCPD have each raised the bar on senior management competence, model risk management, and data protection; leaving firms exposed if they move too fast, and outpaced if they move too slowly.

內容

Understanding the Risks:
- How LLMs work – capabilities, limitations, and why hallucinations occur
- Key risk areas in practice, e.g. accuracy, data privacy, model misuse, etc.
- Hong Kong regulatory expectations

Capturing the Opportunity:
- Building "factories for intellectual work": a practical framework
- Designing AI effective workflows: inputs (data + compute), processes (structured prompts and workflows) and outputs (validated content)
- Collaboration of humans and AI agents

對象

This webinar is designed for compliance officers, in-house counsel, risk managers, manager-in-charge of IT and senior management of SFC-licensed corporations and HKMA-registered institutions.

Technology, operations, and front-office professionals who interact with AI tools in regulated activities will also benefit, regardless of prior technical background.