AI Snips
Chapters
Transcript
Machine Learning Relies On Unpaid Human Work
- Modern machine learning depends heavily on massive, unpaid human labor for labeled data rather than lone genius programmers.
- This data-driven automation treats human-generated input as a free resource, skewing value away from contributors.
Data Is Social, Not Oil
- Major tech platforms show very low labor shares because they treat user-generated data as an untapped resource.
- Calling data “oil” is misleading since data is created by social interactions and should be valued accordingly.
Build Data Cooperatives With Fiduciary Duties
- Create mediators of individual data to collectively bargain, guarantee quality, and train members for digital work.
- Design these organizations with fiduciary duties and no conflicts of interest to protect participants.

