Two free lessons to start

AI security lessons with safe hands-on labs

Start with two free lessons and their cloud labs. Paid access opens the wider learning library.

Prompt Injection Free

Prompt injection is an attack in which untrusted input contains instructions intended to change an AI system's behavior or override the developer's intended rules.

RAG Security Free

RAG security protects retrieval-augmented generation systems from untrusted documents, manipulated retrieval, data leakage, and instructions hidden inside retrieved content.

LIVE · PAID ACCESSChecking access…

Follow an AI workflow, spot where trust breaks, and practice a safer response.

LIVE · PAID ACCESSChecking access…

Look beyond the model and practice protecting the systems that support it.

ON THE HORIZON

A bigger world of learning.

1,020 course concepts across 7 tracks. A glimpse of the questions we want to explore next.

Planned, not published. These lessons and labs have not been built yet. Release dates and inclusion in paid plans are not confirmed. Current purchases cover the live catalog only.

Model resiliencePlanned

COURSE CONCEPT / 001

What would make a confident model get it wrong?

Explore the gap between a convincing prediction and a dependable system.

Lesson + lab conceptNot released

AI application safetyPlanned

COURSE CONCEPT / 001

The response seems helpful. What happens next?

Follow the decisions an AI application makes when trust is misplaced.

Lesson + lab conceptNot released

Agent boundariesPlanned

COURSE CONCEPT / 001

It followed the instruction. It crossed the line.

Explore how a helpful assistant could act beyond the authority it was given.

Lesson + lab conceptNot released

Data & knowledge trustPlanned

COURSE CONCEPT / 001

The answer changed. The question didn’t.

Trace how the information behind an answer can change what a system believes.

Lesson + lab conceptNot released

Platform hardeningPlanned

COURSE CONCEPT / 001

The model is protected. Is everything around it?

Explore the hidden dependencies that can shape an AI system’s safety.

Lesson + lab conceptNot released

Privacy & assurancePlanned

COURSE CONCEPT / 001

You set the rules. Can you show they held?

Explore what it takes to make privacy, oversight, and recovery dependable.

Lesson + lab conceptNot released

Explore hands-on security platforms

Compare ten gamified cybersecurity and AI security learning options, from guided labs to capture-the-flag challenges.

Read the top 10 guide