Participate in practical
AI security research and education
AI War Games welcomes researchers, educators, builders, and security teams who want to learn from labs, run authorized self-directed experiments, contribute scenarios, and help publish practical AI security lessons.
Ways to participate
The project is designed for education, research feedback, and community contribution rather than commercial service delivery.
Research groups
Use the preview to study realistic AI security failure modes and help refine testing methodology.
- ✓Self-directed experimentation
- ✓Authorized systems only
- ✓Optional anonymized observations
Educators and study groups
Use labs to teach prompt injection, data extraction, social engineering, and agent/tool misuse concepts.
- ✓Browser-based lessons
- ✓Guided missions
- ✓Safe classroom discussion prompts
Authorized security teams
Run directional, self-directed checkups for systems you own or are explicitly authorized to evaluate.
- ✓No certification claims
- ✓No managed testing service
- ✓Feedback improves the preview
Scenario contributors
Contribute mission ideas, defense patterns, scoring suggestions, and public learning examples.
- ✓Community scenarios
- ✓Responsible disclosure
- ✓Future public methodology notes
Research-use boundaries
These boundaries keep AI War Games aligned with education, grant-funded research goals, and responsible security practice.
Help shape public AI security learning
We plan to publish educational notes, methodology writeups, and anonymized aggregate lessons as the research preview matures.