Hong Kong Ethical Artificial Intelligence Framework (Version 2.0, December 2025)
Section 5: AI Application Impact Assessment – Hong Kong Ethical Artificial Intelligence Framework (Version 2.0, December 2025)

Hong Kong Ethical Artificial Intelligence Framework (Version 2.0, December 2025) 5.2: Review the assessment at the four lifecycle checkpoints, annually and on major change, and retain copies

What does 5.2 require?

The assessment is reviewed at four key stages (Project Strategy and Planning, Project Development, System Deployment before rollout, and System Operation and Monitoring), regularly such as annually and whenever major changes occur, with copies retained for historical record. At project request the Project Team categorises the project by the gating questions and answers the strategy-stage questions, endorsed by PSC/PAT or, for high-risk projects, the IT Board or CIO; during analysis and design it answers the development-stage questions, updates earlier answers and, where third-party technology or data is involved, completes the third-party questions before procurement, endorsed by PSC/PAT; before rollout it answers the deployment questions, updates the whole assessment and reviews third-party questions, endorsed by PSC/PAT; in maintenance the Maintenance Team updates the assessment and escalates significant issues, the Maintenance Board handles escalations and the IT Board or CIO monitors high-risk projects. Final go or no-go approval for deployment rests with the PSC/PAT, endorsed by the IT Board or CIO as appropriate, on the balance of benefits and mitigated risks.

Maintained by Gerard Blokdyk

Other controls in Section 5: AI Application Impact Assessment – Hong Kong Ethical Artificial Intelligence Framework (Version 2.0, December 2025)

Query this from an agent

The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.