What's New
AI Graders: Improved Answer Recognition
Effective date: Jul 10, 2026
We've updated the base model behind the AI Graders to better handle the full range of valid responses and edge cases across the entire course. This means fewer correct answers being rejected.
What this means for you:
- Valid answers that use alternative phrasing, formatting, or approaches are now more likely to be accepted.
- This update applies to all graded exercises in the course — no action is needed on your part.
- If your answer was previously rejected, you can resubmit it now.
Other updates
More updates will be published in the table below as improvements to the labs continue to be made.
| Date | Lab Module/Challenge Lab | Status |
| 2026-07-10 | Reconnaissance for AI Targets | Updated |
| 2026-07-10 | Assembling The Pieces - Capstone Red Team Engagement | Updated |
| 2026-07-10 | Introduction to Red Teaming Al Systems | Updated |
| 2026-07-10 | Attacking Embeddings | Updated |
| 2026-07-10 | Attacking Model Context Protocol and Tool Surfaces | Updated |
| 2026-07-10 | Supply Chain Attacks on Al/ML Systems | Updated |
| 2026-07-10 | Al Infrastructure and Deployment Exploits | Updated |
| 2026-07-09 | Threat Modeling for Al-Enabled Targets | Updated |
| 2026-07-09 | Attacking Al Agents | Updated |
| 2026-07-09 | Attacking Multi-Agent Systems & Agent-to-Agent (A2A) Protocol | In Progress |
| 2026-07-06 | Exploiting RAG Pipelines | In progress |
| 2026-07-01 | Assembling The Pieces - Capstone Red Team Engagement | Updated |
Important Notes
- These updates are live now — no action is needed to receive them.
- Your existing progress is not affected by these updates.