Skip to content

Instantly share code, notes, and snippets.

A RAI assessment includes stakeholder definitions of direct and indirect types. For example:
---
# Direct Stakeholders
## End user
Who will be evaluated or monitored by the system, whether or not by choice? Who will the system make predictions or recommendations about? E.g., registered customer, patients at X hospital
The HARMS taxonomy lays out the CATEGORY: TYPE OF HARM, and then gives examples of the kind of harm that can be inflicted.
---BEGIN---
## Risk of injury: Physical injury
Technology can cause physical harm.
| Harm | Description | Consideration(s) | Example |
|---|---|---|---|
Pros Cons
You're fully aware how difficult the challenge will be. You may give up because it seems impossible.
You leave nobody behind and have full integrity. You could have been better off leaving somebodies behind.
Being grounded in reality makes it all more real. You can't construct the path forward because reality weighs too heavily.
Pros Cons
You're one of the regular people. Your level of authority is ambiguous.
The noise of the 'people' gets fully heard and you totally get it. You're too deep in the noise and miss the forest from the trees.
If you fail then it won't be a big deal. If you succeed then nobody will really notice.
Pros Cons
It's a classic approach that fits your generation. Folks might not be old-fashioned like you.
Your owning accountability as THE leader. You better be right ... or else.
You can dream really big without constraints. You can easily get ignored by being out of touch.

Keybase proof

I hereby claim:

  • I am johnmaeda on github.
  • I am maeda (https://keybase.io/maeda) on keybase.
  • I have a public key whose fingerprint is 84AF 4067 DAF0 FC8F 0830 7D4F 268C 6C32 B0A2 F196

To claim this, I am signing this object: