Practical Strategies for Successful AI Adoption in your Organisation
summary
Claims AI typically delivers 20% productivity gains but succeeds only with data quality, governance, logging and guardrails; recommends concrete standards like NIST AI Risk Management Framework, ISO 42001 and Microsoft’s Responsible AI checklist, plus a cross‑discipline AI management group. Cites real failures: 2010 market flash crash, Microsoft Tay poisoning, and a federal case where ChatGPT hallucinated citations. Advises transparency to customers (offer opt‑out with ~+20% cost), one hour basic user training, diverse review panels, RAG/agent caution, and continuous monitoring and auditing.
AI-written summary · about
chapters (16)
- 0:00AI adoption challenges
- 3:00AI productivity not job killer
- 6:07AI myths and risks
- 10:43Algorithmic bias mitigation
- 13:15Responsible AI principles
- 18:17Cost speed vs responsible AI
- 23:59AI regulation overview
- 27:13AI standards and frameworks
- 31:27NIST AI risk framework
- 34:13AI risk assessment matrix
- 36:01AI content copyright
- 37:48AI safety explainability security
- 41:47AI management group
- 44:15Why AI adoption struggles
- 46:20AI tool training
- 47:24AI ops model summary
related talks
26mResponsible AI-What, Why, How and Future!
13mAI as a Strategic Asset: Building Competitive Advantage
38mKeynote: Thriving in the Age of AI
48mOperationalizing AI Across the Enterprise
05mPragmatic AI Adopton is Hard!
26mSafeguarding AI: How to Manage AI and its Potential Risks
24mAI Governance for developers: A practical framework for deploying autonomous coding agents safely…
38mBuilding Responsible AI in the Age of GPT