“You Can’t Just Turn It On”: The Hidden Work Behind Enterprise AI
summary
Enterprise AI causes real incidents: an agent deletes production, a Meta agent leaks employee data, and Samsung temporarily bans ChatGPT after staff paste company secrets. Internal behavior metrics are eye‑opening: 58% allowed to use AI, 82% using unmanaged accounts, and 22% paste PII/PCI into GenAI. Regulation (EU AI Act) and buyer demands force five design priorities—visibility, control, predictability, accountability, recovery—and products add granular permissions, moderation rules, audit trails, BYO model options and undo/rollback flows to meet compliance and prevent catastrophic misuse.
AI-written summary · about
chapters (12)
- 0:00Introduction and background
- 3:02AI incidents and risks
- 5:11AI as a UX problem
- 6:57Enterprise AI adoption stats
- 7:39Enterprise requirements for AI
- 9:26EU AI Act and UX responsibility
- 10:29Five UX frameworks for AI
- 13:58Accountability and human oversight
- 15:45Consumer AI regulation trends
- 17:34Building trust in AI products
- 18:16Ten principles for secure AI UX
- 20:42Q&A session
related talks
30mWhy Enterprises Struggle to Adopt AI | Corporate AI Challenges
14mAI Sprawl: The Hidden Risk Threatening Your Enterprise Now
08mWhy AI Governance Matters More Than Ever
28mIs GenAI the technology revolution we need? How can we harness it at work?
33mUnlocking Agentic AI: [...]
24mAI Governance for developers: A practical framework for deploying autonomous coding agents safely…
48mPractical Strategies for Successful AI Adoption in your Organisation
26mSafeguarding AI: How to Manage AI and its Potential Risks