Hong Kong Organisations Boost AI Training as Adoption Reaches 95%
Nearly all Hong Kong organisations now use artificial intelligence in daily operations, with most implementing safeguards to keep humans in control of AI decisions, according to compliance checks by the Office of the Privacy Commissioner for Personal Data.
The PCPD surveyed 60 organisations across 16 sectors in January 2026. Fifty-seven organisations (95%) reported using AI - up 15 percentage points from the previous year.
Human Oversight and Data Protection
Among the 24 organisations that process personal data through AI systems, 19 (79%) adopted a "human-in-the-loop" approach. This means humans retain control over decisions and can catch or correct AI errors before they affect people.
The remaining five organisations used a "human-in-command" model, where staff review AI outputs and intervene only when needed.
All 24 organisations disclosed their use of personal data collection to individuals through formal statements. Seven organisations (29%) specifically mentioned AI tools in these disclosures.
Training and Governance Rise
More than 80% of organisations processing personal data through AI provided AI-related training to employees - an 8-percentage-point increase year-on-year. Eighteen organisations included content on AI privacy risks in their training programs.
Close to 80% established dedicated AI governance structures, such as committees or appointed personnel responsible for overseeing AI systems. Fifteen organisations (63%) referenced the PCPD's AI guidelines when handling personal data.
Security and Testing Standards
All organisations using AI for personal data implemented access controls, encryption, and penetration testing. Five organisations (21%) added AI-specific security alerts and conducted red-teaming drills.
Nearly all organisations tested AI systems before deployment. Twenty-three (96%) ran pre-implementation tests, while 19 (79%) conducted privacy impact assessments beforehand.
Twenty-two organisations (92%) created data breach response plans. Fifteen (63%) conducted regular internal or independent audits.
Generative AI in the Workplace
All 24 organisations permitted employees to use generative AI at work. Seventeen (71%) had written policies to guide proper use, while five (21%) planned to develop them.
The PCPD found no violations of Hong Kong's Personal Data Protection Ordinance during the compliance checks.
What HR Teams Should Know
Ada Chung, Privacy Commissioner for Personal Data, said organisations must address privacy risks as AI adoption accelerates. She recommended developing AI strategies, conducting privacy assessments, implementing appropriate human oversight, and regularly reviewing AI systems' impact on personal data.
For HR professionals, the data shows a clear trend: organisations treating AI governance as a structured discipline - not an afterthought - are investing in employee training and clear policies. This approach protects both the organisation and the people whose data it handles.
HR teams overseeing AI adoption should consider whether your organisation has the governance structures, training programs, and privacy safeguards these leading organisations have put in place. AI for Human Resources resources can help build the foundation. For senior HR leaders, an AI Learning Path for CHROs addresses strategy and governance directly.
Your membership also unlocks:
