Re: thoughts on SOS, participatory agenda setting etc

It seems to me that it would be good if: a) the objectives of each standard were documented in an open, standard, machine-readable format like StratML, and b) the groups specifying them were to document their ethical values in StratML format as well so that others, including AI agents, can evaluate and report on how well their performance reflects their values.
Owen


-----Original Message-----
From: Paola Di Maio <paola.dimaio@gmail.com>
To: public-aikr <public-aikr@w3.org>
Sent: Thu, Mar 7, 2019 11:11 pm
Subject: Fwd: thoughts on SOS, participatory agenda setting etc


Dear AIKR @W3C
After some pondering, I think I have identify a sense of scope  and direction forthe work we could do here, which should be feasible
This  group was started to attempt to address great confusionand lack of orientation in the AI community, in particular in relation to AI ethics (ethics rests on transparency, accountability, understandability etc)
It is becoming clearer to me that KR - a minimal set of explicitly represented concepts and terms - could helpto address at least one problem, the lack of interoperability and lack of accessibilityof the standards being devised around AI ethics from the various standardization bodies
I am therefore sharing a few slides which I created for an IEEE  workgroup to explain what I see the work could be, the focus is the actual KR of the standard itself, but the arguments applyto all the standards and efforts in progress
Part1https://drive.google.com/file/d/1aQxAYjUNYrsf_5qeHqP0O3FGOpermWAx/view
part 2https://drive.google.com/open?id=1i0mYTX2ylyGKu8glmaA-QxInmVVWeeFN
(hope these links work)
sorry if I mumble a bit, may have to re-narrate when I have time
Comments, thoughts?
Can we as a W3C WG aim to capture some common denominator among different standardsand aim to produce a set of metadata to represent AI ethical concerns in some kind ofschematic, implementation independent way?
I will then start drafting a paper as promised
PDM