Human-AI Collaboration Models for Better Decisions and Faster Workflows
summary
Presents concrete human–AI interaction models (centaur, cyborg, ensemble, author-ghost, resonator, agent) and gives criteria for selecting among them based on task risk and complexity, social vs asocial domains, and measurable outcomes. Reports findings that 80% of companies see no organization-level gains and that stronger autonomy can worsen decisions (people “fall asleep at the wheel”), and recommends verification cycles, an autonomy slider, periodic no-AI practice, and tools such as GitHub Copilot and Anthropic agents while noting only 16% prefer pure automation.
AI-written summary · about
chapters (11)
- 0:00Generative AI intro
- 4:21AI productivity debate
- 7:10AI overtrust in drowsy driving
- 10:43Skill loss in aviation
- 14:39AI interaction risks
- 18:50AI social vs antisocial uses
- 22:44AI interaction models
- 27:24Ghostwriter and resonator models
- 30:13Redesigning workflows with AI
- 33:01AI agent predictions 2026
- 35:28Q&A
related talks
20mHuman above AI
24mAI as a Teammate: Redefining Agile for Human–AI Collaboration
47mHuman Robot Agent: Scouting AI Beyond Hype and Doom
13mHuman + Machine
13mHumans, AI, and Mythical Creatures: Are You a Cyborg or a Centaur?
36mStrategies for Human-AI Interaction and Collaboration
27mAgentic Engineering: Working With AI, Not Just Using It
29mThe AI Teammate Is Here — Now What?