香港可信任人工智能產業聯盟Alliance goals
Hong Kong · Industry collaboration
Build AI progresson trust.
We connect industry, technology, policy and public-interest perspectives to make AI accountable, understandable, secure and useful—so governance becomes infrastructure for responsible innovation.
Trust models
Website languages
Alliance base
Alliance goals
Make Hong Kong AItrustworthy.
Treat safety, clear responsibility and public confidence as infrastructure for innovation, market access and industry growth.
Make AI developed and used in Hong Kong more trustworthy through disclosure, human oversight and reviewable evidence.
Advocate clear, evidence-based governance that protects people while enabling responsible innovation and industry growth.
Build four practical reference models covering responsibility, transparency, lifecycle safety and human-centred public value.
Four reference models
Four models fortrustworthy AI.
These in-development models translate recurring governance themes—responsibility, transparency, lifecycle safety and human-centred value—into practical questions. They do not yet represent certification, assurance or an approved technical standard.
Governance & accountability
Named ownership, risk-proportionate oversight, escalation, auditability and responsibilities that cannot be outsourced.
Transparency & evidence
Disclose when AI is involved; explain material limits, data provenance and outputs; preserve the right to know and seek human review.
Safety & resilience
Lifecycle safeguards for data and model custody, access, testing, monitoring, incidents, third parties and operational change.
Human-centred value
Protect rights and welfare, accessibility and meaningful human judgement; test outcomes against real needs and measurable public benefit.
Development path
Progress withits status visible.
Public information distinguishes confirmed work from activity that is still forming or proposed.
Alliance identity established
The public identity and initial programme direction are defined.
ConfirmedProgramme formation
Contribution routes and governance practices are being shaped.
In formationFormalise participation
Publish approved participation and working-group arrangements.
ProposedRelease reviewed outputs
Publish dated, attributable practice notes and updates.
ProposedTrust Lab
Turn principles intoreview questions.
Choose a principle to see the questions it brings into an AI project. This learning aid is not a certification or assurance score.
PRINCIPLE / 01
Who owns the decision?
Accountability begins with named responsibility, an escalation path and a reviewable record.
- Is a decision owner named?
- Which duties remain with the deploying organisation?
- Is evidence retained for review?
News & media
Updates &governance reading.
Alliance updates are separated from independently published commentary. External articles are presented for reference and do not by themselves constitute an Alliance position.
Initial programme framework takes shape
The framework centres on guidance, dialogue, capability and international connection.
Alliance website launches in three languages
Public information is available in English, Traditional Chinese and Simplified Chinese.
Related reading
Independent perspectives on trusted AI
Selected published commentary by Leonard Chan and collaborators. Links open the original publisher pages.
Hong Kong as a global hub for AI governance standards
Jiang Miaomiao & Leonard ChanA standards-hub proposal linking ethics, commercial trust, data governance, human–AI interaction and responsible innovation.
Governing AI disorder and protecting the public interest
Leonard ChanClear responsibility, prohibited boundaries, privacy protection and transparent labelling as foundations for public trust.
Building an AI governance system: Hong Kong has much to contribute
Leonard ChanArgues that safety is infrastructure for innovation and calls for layered governance from models to agents, sectors and rights.
Hong Kong can become a global AI governance testbed
Leonard ChanA human-centred case for the right to know, governance sandboxes and Hong Kong as a bridge between Mainland and international practice.
Participate
Contribute experience.Build common ground.
Industry bodies, enterprises, technology companies, professional services, universities, researchers and public-interest organisations can help shape the programme.
View contact route
Founding relationships and the interim contact route are disclosed on the governance page.