PDF

Index

Cickable topics.

Index

Edit page ⓘ New to GitHub? Get a free GitHub account. Don’t worry if GitHub asks you to create a “fork” — that is simply your editable copy of the project. Make your change, click Propose changes, then Create pull request to send your suggestion to the AI Exchange maintainers for review.

Permalink: https://owaspai.org/go/index

Find clickable topics in alphabetical order below. For an overview of threats and their controls, see the Periodic table of AI security.

A

Adversarial attacks
Agent message structure manipulation
Agentic AI
AI engineering
Alignment

B

Bias

C

Compliance
Continuous validation
Contribute
Controls
Copyright
Cultural sensitivity

D

Data and model governance
Data disclosure in model output
Data poisoning of train/finetune data
Denial of model service
Direct prompt injection

E

EU AI Act
Evasion
Explainability

F

Federated learning

G

GDPR
Generative AI
Governance

H

I

Indirect prompt injection

J

K

L

LLMs
Logging

M

MCP
Media
Model alignment
Model input leak
Model inversion / Membership inference
Model output contains injection
Model poisoning in development-environment
Model poisoning at runtime
Model poisoning through data poisoning of train/finetune data
Model direct leak in runtime
Model poisoning in supply chain
Model direct leak development-time
Model exfiltration
Monitoring

N

O

Oversight

P

Periodic table
Privacy
Prompt injection

Q

R

Red teaming
References
Responsible AI
Risk analysis

S

Safety training
Sponsoring
Supply chain management

T

Testing
Threat modelling
Threats
Training data leaks
Transparency

We are always happy to assist you!

Slack

Send us a message

Have questions about AI security? Want to contribute to our mission? We'd love to hear from you. Reach out through any of our channels or use the contact form.