Professional Information Security Association (PISA)

Unlocking AI Agents: Exploring Opportunities, Risks, and Proprietary Power (11 Mar 2026)

Date: 11 Mar 2026 (Wed)
Time: 7:15pm – 8:45pm
Venue: ADC302, HKU Space Admiralty Centre
Language: Cantonese
Fee: Free
Register: https://bit.ly/pisa260311

Highlight

Topic: AI Agent Era: From Generative AI to Autonomous Workflows — Opportunities & Risks
Speaker: Kwong Chak, Yeung, Founder & CEO, DotAI
• Shift to AI Agents: Technological Evolution
• AI Agents: Multi-Step Tasks & Autonomous Analysis
• Enhanced Efficiency: Task Decomposition & API Integration
• New Risks: Automated Phishing & Misinformation
• Essential Governance: Addressing Cybersecurity Risks
• Cybersecurity: Balancing Innovation & Risk Management
• Evolving Governance: Mitigating Technological Threats

Topic: The Blueprint Session: Mapping Your Path from Public AI to Proprietary Power
Speaker: Andrew Cheung, Co-Founder & CEO, JARY AI | CEO, Upnext Technology (UPNX)
• Why public AI is quietly holding you back
• The anatomy of a real AI Agent
• The AI Expense Trap
• Your Personal AI Blueprint
⁠• Privacy and Security in AI Realm

For detail, please contact us via Email info@pisa.org.hk, Messenger https://m.me/pisahkg or PISA’s WhatsApp Community chat group.

Exploring the Challenges of Generative AI in Intellectual Property Protection (11 Feb 2026)

Date: 11 Feb 2026 (Wed)
Time: 7:15pm – 8:45pm
Venue: ADC303, HKU Space Admiralty Centre
Language: Cantonese
Fee: Free
Register: https://bit.ly/pisa260211

Highlight

With the rapid development of artificial intelligence (AI) technology, the legal framework for copyright faces new challenges. This training aims to help participants gain a deeper understanding of copyright issues and legal responsibilities in the age of AI, while providing best practices for protecting intellectual property rights.
Session detail:
– What are the potential legal issues associated with Generative AI?
– Who owns the copyright for generative texts and music clips?
– Are there any notable copyright dispute cases related to artificial intelligence?
– Since artificial intelligence often involves cross-boarder disputes, how can they be effectively addressed?

Speaker: Professor Albert So
Professor Albert So is a practicing solicitor of High Court of Hong Kong and the Chairman of Certified Anti-Money Laundering Professionals Organization. He is also the Chairperson of Hong Kong Mediation and Arbitration Centre, Chairman of Wealth Succession Planning Association and President of Law Society of One Belt One Road. He is currently university lecturer and programme director of various law and mediation programmes of universities in Hong Kong and overseas.
Professor So obtained his bachelor’s degree in laws at The University of London, postgraduate studies in laws at both The University of Hong Kong and The King’s College London. He received his professional postgraduate education in civil law, business and finance at The Tsinghua University, University of Cambridge and Harvard University (John F. Kennedy School of Government).
Before joining the legal profession, Professor So had been one of the prosecution team members of regulatory authority, responsible for conducting criminal investigation, handling litigation and disciplinary-related professional proceedings. Throughout his professional time as a regulator, he had personally handled over 800 financial disputes and their relevant litigations.

For detail, please contact us via Email info@pisa.org.hk, Messenger https://m.me/pisahkg or PISA’s WhatsApp Community chat group.

Supporting Event: AI Security Workshop – part 2 (31 Jan 2026)

Date: 29 Nov 2025 (Sat)
Time: 9:30am – 12:30pm
Venue: FTC 1407, HKUSPACE Campus (North Point)
Fee: Free (Pre-registration is required)
Register: https://dracosec.tech/en/form/offsec (Laptop required)

Highlight

OffSec HK Chapter – AI Security Workshop
Adversarial AI: Discover, Defend, Deliver – Part 2

In an era where AI drives applications from autonomous vehicles to fraud detection systems, even minor manipulations can compromise model integrity. Consider scenarios such as deceiving facial recognition systems with subtle stickers or corrupting recommendation algorithms to propagate false information. This introduces the domain of Adversarial Machine Learning (AML), where vulnerabilities hide within neural network architectures.

This workshop offers a practical, hands-on exploration of these challenges. It clarifies AML risks and provides actionable methods for developing resilient AI solutions. Informed by recent advancements and actual events, such as the 2023 Uber autonomous driving vulnerabilities and the surge in 2025 LLM prompt injection attacks. This session covers the complete AML process: identifying concealed flaws, implementing effective defenses, and deploying fortified models in production environments.

Note: Please bring along your laptop computer for hands-on tasks.

Conducted in: Cantonese (Spoken) with English materials

For detail, please visit https://dracosec.tech/en/form/offsec.

Beyond the Prompt: Architecting Secure, Enterprise-Grade AI with Google’s Agentic Frameworks (30 Jan 2026)

Date: 30 Jan 2026 (Fri)
Time: 7:15pm – 8:45pm
Venue: ADC315, HKU Space Admiralty Centre
Language: Cantonese
Fee: Free
Register: https://bit.ly/pisa260130

Highlight

Join us for a technical deep dive into the next evolution of artificial intelligence: agentic AI. This session, led by Google Cloud’s Gen AI Specialist Jim Ng, moves beyond single-prompt interactions to explore how autonomous AI agents can reason, plan, and utilize tools to automate complex, multi-step business processes.
This talk will demystify the architecture required to build and deploy sophisticated AI agents at an enterprise scale. We will introduce Google’s “Enterprise Agentic Foundation,” explaining how components like the Vertex AI Agent Engine, Agentspace, and the open Agent2Agent (A2A) protocol simplify the development and management of these systems.
With a special focus on the needs of security professionals, we will explore how platforms like Apigee provide essential governance and control over LLM APIs. Key security topics will include:
– Enforcing security policies, rate limiting, and metering for API endpoints.
– Redacting and masking sensitive data in responses.
– Inspecting prompts for malicious content and jailbreak attempts using Model Armor.
– Securely routing requests across multiple models based on defined criteria.
Attendees will leave with a clear, architectural understanding of how to build, manage, and secure the next generation of agentic AI solutions.

Speaker: Jim Ng
Jim Ng is a Customer Solutions Architect at Google Cloud, where he specializes in designing cutting-edge solutions for customers. He holds an advanced L400 certification in AI/ML and an L300 certification in Application Modernization, underscoring his deep technical expertise. He has a proven track record of supporting major Hong Kong Google Cloud customers in designing, implementing, and deploying AI and Data systems.
A proud alumnus of the Chinese University of Hong Kong, Jim graduated with first-class honours from the Faculty of Engineering. He is passionate about sharing his knowledge and is excited to introduce the members of the Professional Information Security Association (PISA) to the transformative power of Google’s Gen AI and Agentic Frameworks.

For detail, please contact us via Email info@pisa.org.hk, Messenger https://m.me/pisahkg or PISA’s WhatsApp Community chat group.