20 of the most operationally relevant terms from the full 50-term lexicon. Download the PDF for the complete set.
Adversarial AI
Techniques and attacks used to manipulate AI systems, causing them to make incorrect or unintended predictions or decisions by exploiting vulnerabilities in models.
Agentic AI
A category of AI systems capable of independently making decisions, interacting with their environment, and optimizing processes without direct human intervention.
AI Agent
A system that autonomously perceives its environment, decides what to do, and takes actions to achieve its goals.
AI Drift / Decay
The tendency for an AI model's performance to degrade over time when deployed in real-world settings with conditions that differ from training.
AI Governance
The set of organizational policies, rules, frameworks, roles, and oversight processes that direct how AI is adopted, developed, deployed, and monitored.
AI Hallucination
A phenomenon when AI produces output that is erroneous or flawed but is still presented in a convincing narrative or form.
AI Lifecycle
The set of phases an AI system goes through: plan and design, collect and process data, build and use model, verify and validate, deploy and use, and operate and monitor.
AI Risk Assessment
A risk-management process for identifying, estimating, and prioritizing risks arising from the operation and use of an AI system.
Foundation Models
Large machine learning models trained on vast amounts of raw and unlabeled data through unsupervised learning that can be adapted to versatile downstream tasks.
Generative AI
The class of AI that emulates the structure and characteristics of input data in order to generate derived synthetic content including images, video, audio, and text.
Guardrails
Layered safeguards to prevent access to bad information and behavior in an AI system, encompassing policies, technical controls, and monitoring mechanisms.
Human-in-the-Loop (HITL)
A risk-control approach for AI where a human is integrated within the AI's decision-making process.
Large Language Model (LLM)
A subset of machine learning that uses algorithms trained on large amounts of data to recognize patterns and respond to user requests in natural language.
Model Risk
The potential for adverse consequences from decisions based on incorrect or misused model outputs and reports, including aggregate risk from model interactions.
Prompt Injection
An attack on an AI system that exploits how an application combines untrusted input with a prompt written by a higher-trust party, causing the system to follow untrusted instructions.
RAG (Retrieval Augmented Generation)
A generative AI system in which a model is paired with a separate information retrieval system. Based on a user query, the system identifies relevant information and provides it to the model in context.
Responsible AI
Conscientious design, deployment, and governance of AI systems aligned with ethical principles, societal values, and legal requirements.
Synthetic Data
Data generated using a purpose-built mathematical model or algorithm that is statistically realistic but artificial, used for model development and training.
Third-Party AI Risk
Risk that arises when an organization relies on another entity to develop, provide, host, operate, or support AI systems or key AI components.
Traditional AI
Also referred to as symbolic or rule-based AI, a subset of AI that performs discrete, preset tasks using predetermined algorithms and rules.
