Control · C-057 · Governance & Accountability
Shadow AI discovery
Network, procurement, expense and software records are reviewed on a set cycle to find AI tools and features in use that are absent from the inventory.
- Control type
- Detective
- Risks addressed
- 3
- Audience
- Heads-of · Practitioner
- Frameworks
- Principle 5 · GOVERN 1
- Responsible AI domain
- Lifecycle stage
- Reviewed
- 2026-11-10
What it does
It finds the AI the organisation did not know it had. Independent data sources are searched for signs of AI use, the results are compared with the inventory, and each discovery is either brought into governance or stopped. Security owns the technical monitoring and the governance function owns the reconciliation. It runs continuously for network data and on a set cycle, at least quarterly, for the rest.
How to implement it
- Monitor outbound traffic for known AI providers and for browser extensions.
- Review expense claims and purchasing card data for AI subscriptions.
- Scan software and cloud records for AI components and enabled features.
- Review supplier release notes for new AI features in products already in use.
- Ask staff, by survey, what they use and why.
- For each finding, decide: approve and register, replace with an approved tool, or block.
Evidence it operates
Design: a discovery procedure listing sources, frequency and owners. Operation: discovery reports; inventory entries created as a result; records of tools blocked or replaced; a trend in the number of unregistered tools found.
Where it fits
Runs in operation as the check on intake. With the AI inventory it is a primary control for incomplete AI inventory, and with the AI policy for shadow AI use. What it finds is also evidence of unmet need, which should shape the approved toolset.
Limits
It cannot see use on personal devices and personal accounts outside the organisation's network. Providers and tools change faster than any list of them. Discovery used only to block and discipline teaches staff to hide; used to understand need, it reduces the problem at source. Monitoring of staff activity has its own legal limits.
Risks this mitigates
Primary mappings
- Incomplete AI inventory High
- Finds systems in use that the record lacks. The organisation cannot list the AI systems and models it builds, buys or uses, so risk assessment, oversight and regulatory duties are applied to only part of what it has.
- Shadow AI use High
- Finds unapproved tools from network, procurement and expense data. Staff use AI tools that have not been approved, placing organisational data and decisions outside any control.
Supporting mappings
- Confidential information disclosed to AI tools High
- Finds unapproved tools where the policy is bypassed. Trade secrets, legally privileged material or information held in confidence for others are entered into AI tools and leave the organisation's control.
Framework and clause references
| Framework | Clause | Title |
|---|---|---|
| Code of Practice for the Cyber Security of AI | Principle 5 | Identify, track and protect your assets Identification of assets. |
| NIST AI Risk Management Framework (AI RMF 1.0) | GOVERN 1 | Policies, processes and practices for AI risk are in place and effective Inventory completeness. |
Related terms
How to cite
Axiom Verity (2026). Shadow AI discovery. Reference Library, C-057. https://axiom-verity.com/controls/shadow-ai-discovery. Reviewed 2026-11-10.
Found an error or a better source? Report a correction.