Glossary
AI governance glossary
The terms that come up when an organisation governs its AI, each defined in one sentence and grouped by the stage it belongs to: discover, assess, control, comply.
10 terms
Across the whole loop
Two terms cover all four stages.
AI control layer
An AI control layer is the part of AI governance that sits where AI is used and applies the rules: it sees every AI system in use and enforces the organisation's policies on each prompt and each agent action.
AI governance
AI governance is the set of policies, roles and controls an organisation uses to decide which AI systems may be used, for what and with which data, and to prove that those decisions are followed.
Terms by stage
Discover
Every AI system in use, shadow AI and personal accounts included, in one continuously synced inventory.
Assess
Every system risk-scored on eight dimensions, one governance score, and the actions that raise it.
Control
Your rules, enforced at runtime on every AI in use, shadow AI included. Written by the business, monitored first.
Comply
EU AI Act classification, continuous control evidence, and exportable reports for your DPO.
Discover
Shadow AI
Shadow AI is the use of AI tools, models or agents for work without the approval or oversight of the organisation's IT, security or compliance teams.
Discover
AI inventory
An AI inventory is a complete, current record of the AI systems an organisation uses, with an owner and a risk classification for each one.
Discover
Agent sprawl
Agent sprawl is the uncontrolled growth of AI agents built across an organisation, often on low-code platforms, with no central record of who owns them, what they can access or who uses them.
Assess
High-risk AI system
Under the EU AI Act, a high-risk AI system is one used in an area listed in Annex III, such as hiring, credit scoring or access to essential services, or one that is a safety component of a product regulated under Annex I.
Control
Runtime enforcement
Runtime enforcement means applying an AI policy to each request as it happens, for example by blocking, redacting or warning before data reaches an AI tool, instead of relying on people to remember a written policy.
Control
Model Context Protocol (MCP)
The Model Context Protocol (MCP) is an open standard for connecting AI applications and agents to external tools and data sources, such as file systems, databases and SaaS APIs, through one common interface.
Comply
EU AI Act
The EU AI Act, Regulation (EU) 2024/1689, is the European Union's law on artificial intelligence, which sorts AI systems by risk, bans some practices outright and sets obligations both for the organisations that build AI systems and for those that use them.
Comply
AI literacy (Article 4)
AI literacy is the skills, knowledge and understanding people need to use AI systems in an informed way, aware of both their opportunities and their risks.
See Qastel in action.
A 30-minute demo of Qastel discovering, assessing and governing AI systems.