Guiding AI Governance Evolution: Traditional to Generative AI
summary
Claim: 86% of Swedish CEOs plan to deploy generative AI within 12 months and foundation models are often built by tech giants scraping web-scale data then trained for long runs. Risks include amplified fairness and privacy harms plus new failures like hallucinations and data leakage illustrated by a phone number leak from a chatbot. Practical actions are to move model approval up the org, update risk frameworks, add MLOps monitoring for toxic language and models that take more than 30 days to train, and tighten vendor due diligence.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 3:43Foundation models overview
- 6:30AI governance core
- 8:57EU AI Act and risks
- 10:22Traditional risks amplified
- 14:01AI governance actions
- 19:23Balancing risks and opportunities
- 20:31Q&A
related talks
17mGenerative AI: Keeping it real
34mGenerative AI Landscape Building a Modern Data Strategy
41mThe AI boom: what and why?
18mNavigating the growth frontiers of generative AI
42mFrom theory to practice: The rise of Generative AI
01h 14mLatest trends in AI
22mBuilding Governance To Maximise AI Business Impact
15mPractical Evolution of Data Science with GenAI