AI glossary
Clear, consumer-friendly definitions of AI and compliance terms.
A
- Adverse Action Notice
- A notice some laws require when a business takes a negative action against you, such as denying credit — including when AI is involved.
- Algorithmic Bias
- When an AI system produces systematically unfair outcomes for certain groups.
- Artificial Intelligence (AI)
- Software that performs tasks typically needing human intelligence, like recognizing images, understanding language, or making predictions.
- Automated Decision-Making (ADM)
- When a computer system makes or significantly influences a decision about you.
C
- Conformity Assessment
- A formal process to verify that an AI system meets legal requirements before it is placed on the market.
- Consent
- Your freely-given, specific, informed agreement — often required before an organization can use AI with your personal data.
D
- Data Minimization
- A privacy principle that says organizations should only collect and use the minimum personal data needed.
- Deepfake
- A very realistic AI-generated image, audio, or video that appears to show a real person doing or saying something they did not.
E
- Explainability
- The ability to understand why an AI system produced a particular output.
F
- Foundation Model
- A large, general-purpose AI model trained on broad data that can be adapted to many tasks.
G
- Generative AI
- AI that creates new content — text, images, audio, or video — rather than only analyzing existing data.
H
- High-Risk AI
- AI used in a context where mistakes can cause significant harm — such as healthcare, credit, hiring, or public safety. Regulators require extra safeguards for these systems.
- Human Oversight
- Requirements that a human can supervise, intervene, or override an AI system's decision.
L
- Large Language Model (LLM)
- An AI model trained on huge amounts of text to generate and understand language.
M
- Machine Learning
- A type of AI that learns patterns from examples instead of being explicitly programmed.
- Model Card
- A short document describing an AI model's purpose, training data, performance, and limitations.
R
- Red-Teaming
- Deliberately probing an AI system to find safety, security, or bias weaknesses before it is deployed.
S
- Synthetic Media
- Images, audio, video, or text created or substantially altered by AI.
T
- Training Data
- The data used to teach an AI system patterns and behaviors. The quality and composition of training data strongly shape the system's behavior.
W
- Watermark
- A mark embedded in AI-generated content to help identify it as synthetic.