AI Engineer Ethics: Shaping Responsible Generative AI
summary
Shows concrete defenses for generative AI: deploy Llama and Falcon endpoints behind request filters (keyword rules, business rules or a classifier like toxygene bird from Hugging Face), use prompt engineering (prepend instructions or tack on tokens to reduce bias), and implement retrieval-augmented generation to fetch live sources (example: query → web search → confirm “Rishi Sunak” to avoid a hallucinated PM). Demonstrations reduce hallucinations and bias; monitoring via thumbs-up/down closes the loop.
AI-written summary · about
chapters (9)
- 0:00AI Engineer Ethics intro
- 3:20Generative AI challenges
- 8:15AI engineer's role
- 10:48Building AI with LLMs
- 14:26Filtering AI content
- 18:48Prompt engineering bias
- 29:18Retrieval augmented generation for accuracy
- 35:19Responsible AI lessons
- 37:17Q&A on governance
related talks
28mEthical AI?!
44mGenerative AI Security — A Practical Guide to Securing Your AI Application
35mAI ethics for tech writers
55mAI Showdown: The Good, The Glitchy, and The Risky
23mTrustworthy Generative AI: The Role of Observability and Guardrails.
40mCode is Life - Building a Responsible Generative AI Solution
45mOverview of Risks, Vulnerabilities and Paths to Safer Use of Generative AI
37mInnovate Responsibly with Generative AI