Open Source Tools To Empower Ethical and Robust AI Systems
Talkat Open Source Summit Europe 2025 · on Monday, 2025-08-25· at 13:30· Elicium 1 (Elicium Level 1)·
summary
A curated GitHub repo with a mermaid taxonomy and runnable Colab notebooks maps open-source tooling across pre‑deployment, runtime and knowledge phases; examples include LLMGuard (Python input/output filters that obfuscate and later restore PII but break streaming), Guardrails AI (validator hub with remote‑judge support), Garak and Promptfoo for probe‑based LLM evaluation, Giskard’s RAG test‑suite generator, ART for adversarial attacks on classic ML, and AI Fairness 360 for bias mitigation — most tools are company‑backed, interoperable, and cover the full ML lifecycle.
AI-written summary · about
chapters (12)
- 0:00Open source AI security tools
- 2:28Rising AI risks and guardrails
- 5:42Taxonomy of AI security tools
- 8:55Runtime filters for LLM input and output
- 12:46Evaluating LLM security with Garak and Promptfoo
- 16:41Giskard for vulnerability scanning
- 18:30Purple Llama toolkit for LLM security
- 21:17Adversarial Robustness Toolbox
- 24:07AI Fairness 360 for bias detection
- 27:17OASP GenAI Security Project
- 28:44FINOS AI Governance Framework
- 31:35Key insights on AI security tools
related talks
26mOpen Source Tools To Empower Ethical and Robust AI Systems
26mSponsored Session: Navigating Open Source AI with Guardrails
11mKeynote: Trusted AI: Bringing Trust Back into AI through Open Source
46mCan you trust your AI?
39mThe Search for Transparency and Accountability in the Age of...
01h 16mLifting the Hood on Big AI - The Future of Transparency and Accountability in AI
43mSecurity of GenAI & LLMs: A New Era of Ethical Hacking
26mBuilding ethical AI: Transparency, bias mitigation, and trust