Generative AI at Work / Unit 8 GAI 190 worked example, 60 to 90 seconds
Asking the Datasheet Questions
The idea
When your work relies on a dataset, from a vendor's training data to the document library behind an internal assistant, Gebru and colleagues' question groups make a useful checklist even without a published datasheet.
In practice
Step one: motivation: who assembled this data, and for what purpose? Step two: composition: what does it contain, how current is it, and does it include personal or sensitive information? Step three: collection and cleaning: how did it get here, and what was removed or changed along the way? Step four: uses: what is it suited for, and what should it not be used for? Step five: maintenance: who keeps it current, and how will you know when it changes?
Where people slip
Questions nobody can answer are findings in themselves. A grounded tool reading from a library with no owner and no update process will drift without anyone noticing.
Your place on the trail
Every lesson rests on earlier ones and opens later ones. Tap any to jump.
Unlocks
- Nothing yet depends on this.
Something wrong in this lesson? Flag it · Sources for Generative AI at Work