30 Terms

AI Glossary

Every enterprise AI term you need to know — defined clearly with practical context.

Free Assessment

How Exposed Is Your Company?

Most companies already have employees using AI. The question is whether that's happening safely. Take 2 minutes to find out.

You get

A short report showing where your biggest AI risks are right now.

A

Agentic AI

Artificial intelligence systems that can autonomously plan, execute multi-step workflows, and take actions.

AI Alignment

The process of ensuring an AI model's goals and behaviors match human values and corporate policies.

AI Bias

Systematic and unfair prejudice in AI outputs, resulting from flawed training data or algorithmic design.

AI FinOps

Operational cost governance for AI usage, including budgeting, tracking, and optimization.

AI Governance

The policies, controls, and operating practices used to manage AI usage safely at scale.

AI Hallucination

A phenomenon where an AI model confidently generates false or fabricated information.

AI Incident Response

A structured process for handling high-risk AI events and policy violations.

AI Observability

The continuous monitoring and analysis of an AI system's health, performance, and outputs in production.

AI Risk

Potential negative outcomes from AI usage, including policy, privacy, financial, and operational impacts.

AI Supply Chain Risk

The hidden security and compliance vulnerabilities introduced by relying on third-party AI models and datasets.

AI Transparency

The degree to which an AI system's operations, training data, and decision-making processes are visible and understandable to stakeholders.

Audit Trails

Traceable records of AI activity, governance actions, and control events.

D

Data Poisoning

A cyberattack where malicious data is deliberately injected into a model's training set to corrupt its behavior.

Department Budgets

Team-level spending controls used to manage AI usage across an organization.

F

Fine-Tuning

The process of retraining a pre-trained model on specialized data to improve specific task performance.

Foundation Model

A massive AI model trained on vast amounts of data, adaptable to a wide range of tasks.

I

Inference Cost

The financial cost incurred each time an AI model processes a prompt and generates a response.

K

Knowledge Grounding

Using approved internal context to improve response relevance in AI workflows.

M

Model Drift

The degradation of an AI model's performance and accuracy over time due to changing real-world data.

Model Governance

Policies that control model availability and usage behavior by team and context.

P

Policy Guardrails

Control checks that evaluate AI inputs and outputs against organization policy.

Prompt Injection

A cyberattack where malicious instructions are hidden within a prompt to manipulate an AI model.

R

Red Teaming (AI)

The adversarial practice of aggressively testing an AI system to discover security flaws, biases, and vulnerabilities.

Responsible AI

A framework for developing and deploying AI in a way that is ethical, transparent, and legally compliant.

Retention Controls

Configurable settings that define how long AI interaction data is stored and who can access it.

Retrieval-Augmented Generation (RAG)

A method where AI responses are informed by retrieved reference content.

Role-Based Access Control

An access model that grants permissions according to organizational role and responsibility.

S

Sensitive Data Protection

Controls that reduce accidental disclosure of confidential data in AI workflows.

Shadow AI

The unsanctioned use of artificial intelligence tools by employees without IT approval or oversight.

U

Usage Analytics

Operational reporting on AI adoption, policy events, and spending trends.

Free Resource

The 1-Page AI Safety Sheet

Print this, pin it next to every screen. 10 rules your team should follow every time they use AI at work.