Open Source Tools To Empower Ethical and Robust AI Systems
Talkat Cloud Native + Kubernetes AI Day Europe 2025 · on Tuesday, 2025-04-01· at 16:30· Level 1 | Hall Entrance N10 | Room G·
summary
Shows a hands‑on survey of open source tools with examples and artifacts: Garak and Gizkar for automated safety probes, LLM Guard and LLM Canary for runtime prompt blocking, Promptfoo for prompt tests, Counterfit/Adversarial Robustness Toolkit and BrokenHill for attacks (including a database reconstruction attack and a greedy coordinate jailbreak demo), AI Fairness 360/Audit AI for bias metrics, notebooks and a deployable runtime AI proxy reference architecture with logs and integrations.
AI-written summary · about
chapters (15)
- 0:00Open source AI tools
- 2:31Robust and ethical AI
- 3:59Evaluating AI models
- 7:52Security with Garak
- 8:59Runtime monitoring with LM Guard
- 10:46Adversarial Robustness Toolbox
- 13:39AI Fairness for bias
- 15:51Nemo Guardrails
- 18:21Lancet for LLM monitoring
- 19:29Inspect AI benchmarking
- 20:34Broken Hill jailbreak testing
- 21:36Other tools and architectures
- 23:20Finos AI governance
- 24:27Key takeaways
- 24:48Q&A
related talks
35mOpen Source Tools To Empower Ethical and Robust AI Systems
11mKeynote: Trusted AI: Bringing Trust Back into AI through Open Source
26mSponsored Session: Navigating Open Source AI with Guardrails
53mHow to Break AI Systems (Before Someone Else Does)
19mKeynote: Towards Building Safe and Secure Agentic AI
46mCan you trust your AI?
24mHow to Build Trustworthy AI
01h 24mOffensive Cybersecurity by AI Promises and Pitfalls