Human-centered AI
summary
Argues that traditional user-centered design is insufficient and proposes community-centered design to catch stakeholders beyond direct users; warns large language models are largely trained on Western internet data embedding cultural bias; gives a rough grade of B− for current industry efforts and notes resources to build large models cost tens or hundreds of millions; recommends interdisciplinary teams and to embed social scientists in product teams early, and lists investments in neuroscience and common-sense reasoning as concrete remedies.
AI-written summary · about
chapters (8)
- 0:00Human-centered AI intro
- 2:10Beyond user-centered design
- 4:36AI benefiting humanity now
- 6:41Democratizing AI resources
- 8:04Bias in AI models
- 10:31Diversity gaps in AI teams
- 12:38Responsible AI recommendations
- 14:48Stanford’s AI research
related talks
24mA Call for Human-Centered AI
31mA more human-centered approach to AI
54mEthical AI Development: Leveraging Human Centered Design for Inclusive and Trustworthy Solutions
42mBrand Humanizing: A humane turn on AI
45mModeling as Mind Redesign: Why participatory AI is better
37mFinding the Problem to Solve: Human-centered AI
23mUser Experience Now Includes Machine Experience
19mMoral Outsourcing: Humanity in the age of AI