The Four Horsemen of AI

I came across a tweet by Dan McQuillan today with a good concept, usable for arguing for SSNA (@nadia: “a human-powered alternative to AI”).

The four horsemen of AI:

  • Fairness
  • Transparency
  • Accountability
  • Interpretability

It is not perfect, because the Horsemen are supposed to be negative forces. I guess McQuillan means that these are negative forces for AI itself, but the rhetoric would be more powerful if it were reversed: unfairness, opacity, etc. I will leave this to @hugi and Nadia. :slight_smile:

3 Likes

I could fill in on my old ideas of human-directed AI. Aimed at the same issues.

1 Like

@asimong sure, go right ahead!

Curious to have this subject more developped, as I’m passionated about AI

Sorry the thread slipped between my fingers :slight_smile:

I’d like to say something much shorter and more helpful, but here is my PhD which ended up being all about this…

http://www.simongrant.org/pubs/thesis/

Please, please, feel free to dip the smallest toe in and then ask me to expand on anything of interest.
I guess, start at the last chapter.