INDEPENDENT ANALYSIS · AV AND INTELLIGENT BUILDINGS
AI governance for AV and intelligent spaces
AI now runs in cameras, microphones, DSPs, room systems, building sensors and cloud-connected AV workflows. Governance keeps those systems transparent, secure and accountable across their lifecycle.
Six principles of AI governance
The principles behind HCAIG-1. Each is put into practice by one or more of its six clauses.
Transparency
People know which AI systems are in a space and what they infer.
In HCAIG-1: Environment Mapping · Behaviour Assessment
Accountability
A named owner for every deployment, and a person who can step in.
In HCAIG-1: Human-in-the-Middle Controls · Continuous Governance & Assurance
Compliance
Meeting the EU AI Act and aligning with ISO/IEC 42001, with evidence to show it.
In HCAIG-1: Regulatory Compliance & Conformity
Ethical use
Systems respect the dignity and autonomy of everyone in the space, not just their users.
In HCAIG-1: Risk & Impact Alignment
Data protection
Audio, video and sensor data is collected only as needed, and protected throughout.
In HCAIG-1: Environment Mapping, alongside UK GDPR
Risk management
Risks are assessed before deployment and again whenever the space or the system changes.
In HCAIG-1: Risk & Impact Alignment · Continuous Governance & Assurance
Governance across the AV value chain
Manufacturers
Document what AI features infer, provide controls to switch them off, and give customers the evidence their compliance teams will ask for.
Integrators
Configure AI features deliberately, record what was enabled and why, and answer governance questions in bids and at handover.
Enterprises and end users
Own the deployment: name an accountable person, tell people what runs in their spaces, and question suppliers before buying.
Frameworks and standards
The standards that shape AI governance and security for AV. HCAIG-1 is being mapped against ISO/IEC 42001 and the EU AI Act; mappings are published once verified.
ISO/IEC 42001
The international standard for an AI management system: how an organisation governs, assesses and improves its use of AI.
For AV: the management backbone for AI features in products and estates.
EU AI Act
The EU’s risk-based AI law: banned practices, plus obligations for high-risk systems and the organisations that deploy them.
For AV: bans emotion recognition in workplaces and education; sets duties for deployers.
NIST AI RMF
A voluntary US framework for managing AI risk, organised around govern, map, measure and manage.
For AV: a common language with US clients and manufacturers.
ISO/IEC 27001 / 27002
The international standards for information security management and controls.
For AV: securing networked devices, platforms and the data they handle.
AVIXA standards
Industry standards for AV system performance and design.
For AV: the technical baseline that governance builds on.
Governance in practice
How governance flows through a real space, and what each step delivers.
1 · MAP
Know the space
Outcome: a full inventory of systems and the people they affect
2 · ASSESS
Know what it infers
Outcome: no hidden inferences about people
3 · CONTROL
Keep a human in charge
Outcome: a tested override with a named person
4 · ALIGN
Weigh the impact
Outcome: risks understood before go-live
5 · ASSURE
Keep it governed
Outcome: re-checked when anything changes
6 · EVIDENCE
Show compliance
Outcome: proof for clients, auditors and regulators
Research agenda
IN PROGRESS
Why ISO/IEC 42001 will reshape enterprise AV governance
IN PROGRESS
Governance gaps in unified communications platforms
IN PROGRESS
Smart-building AI: the new frontier for cyber-physical risk
PUBLISHING DECEMBER 2026
Get the Governing the Room briefing paper
The candidate governance gaps when AI moves into meeting rooms and buildings, and what the EU AI Act says about each.
Occasional updates only. Unsubscribe anytime. Sign-ups are managed by WordPress.com.
Independent insight
Vendor-neutral and standards-aligned. Published by Benedict Onodu, author of HCAIG-1.
© 2026 Benedict Onodu · The AV Governance Report · Contact
