Ploba logo
Discover
Hackathons
People
Learn
AI Digest
Search
⌘K
Sign in
Get started
AI glossary
Safety
AI reliability, security, alignment, misuse prevention and responsible development.
Search the glossary
All
Agents
16
Development
29
Fundamentals
10
Generative AI
18
Machine Learning
24
Models
14
Protocols
12
Safety
12
Safety terms
12 terms
AI Alignment
The effort to ensure an AI system’s behavior and objectives reliably match intended human goals and constraints.
AI Bias
Systematic skew in AI outputs resulting from flawed data, design, or deployment choices.
AI Guardrails
Safety mechanisms and policies put in place to ensure AI systems operate within defined ethical, legal, and operational boundaries.
AI Jailbreaking
The practice of crafting specific prompts to bypass an AI model's built-in safety restrictions and moderation filters.
AI Red Teaming
Structured adversarial testing intended to uncover unsafe behavior, security weaknesses, and failure modes in AI systems.
AI Safety
The field of research and engineering focused on preventing AI systems from causing harm, whether through mistakes, misuse or behaviour that drifts from what people intended.
AI Watermarking
Marking AI-generated images, video, audio or text, visibly or invisibly, so people and software can tell it was made by AI.
Constitutional AI
A training method in which an AI model critiques and revises its own answers against a written set of principles, reducing the need for humans to label harmful content.
Data Poisoning
An attack that corrupts the data a model learns from, so the finished model behaves wrongly or contains a hidden trigger planted by the attacker.
Deepfake
Realistic AI-generated or AI-altered video, audio or images that show a real person saying or doing something they never did.
Prompt Injection
A security vulnerability where malicious instructions disguised as normal input manipulate an AI model's intended behavior.
Sycophancy
When an AI model tells people what they want to hear, agreeing with them or flattering them, even when that makes its answer less accurate.