Re: DRAFT (Late) Feedback for NIST on the AI Standards Workshop

The US AI plan is now available in StratML format at 
http://stratml.us/drybridge/index.htm#USLAI Here are some comments that 
are patently obvious, at least to me:

    With reference to Goal 2: Trustworthiness
    <http://stratml.us/carmel/iso/USLAIwStyle.xml#_1f31f0fc-c831-11e9-b146-73880edd4871>,
    in groups exceeding Dunbar's Number, President Reagan's admonition
    applies:  Trust but verify.  U.S. federal agencies can hardly be
    deemed trustworthy if they fail to comply with the very law that
    requires them to report their performance in open, standard,
    machine-readable format.  To do so would be a fools errand, i.e.,
    choosing to remain blissfully ignorant.  While that may be prudent
    in dictatorships, it would belie the avowed nature of the "land of
    the free and the home of the brave."

    Objective 2.2: Performance Metrics & Data Sets
    <http://stratml.us/carmel/iso/USLAIwStyle.xml#_1f31f21e-c831-11e9-b146-73880edd4871>
    points in the right direction but should also note that metrics and
    data sets should be published in open, standard, machine-readable
    formats.

    Goal 3: Partnerships
    <http://stratml.us/carmel/iso/USLAIwStyle.xml#_1f31f3e0-c831-11e9-b146-73880edd4871>
    would be well served if each of the prospective partners were to
    render their own plans in StratML format and apply the
    stratml:Relationship elements to strategically align,
    cross-reference, and link their common and complementary objectives
    to each other.

    Likewise Goal 4: International Engagement
    <http://stratml.us/carmel/iso/USLAIwStyle.xml#_1f31f8e0-c831-11e9-b146-73880edd4871>
    would be well served if nations with common values were to engage
    each other as partners via StratML-enabled services.

    Doing so would semi-automatically address Objective 4.3
    <http://stratml.us/carmel/iso/USLAIwStyle.xml#_1f31fb56-c831-11e9-b146-73880edd4871>
    by enabling value-added intermediaries to track and report progress
    on the development of AI standards.

    With reference to Objective 1.1: Standards Coordinator
    <http://stratml.us/carmel/iso/USLAIwStyle.xml#_1f31edbe-c831-11e9-b146-73880edd4871>,
    it may be fine to appoint one but not without giving her the
    StratML-enabled tools to do her job.  The proliferation of CXOs
    calls to mind the politically incorrect phrase "too many chiefs and
    not enough Indians." Creating scapegoats is an all-too-common bad
    practice of governance.  No one can succeed in an environment fails
    to support the necessary means.  Another relevant truism is that the
    mission of a good manager is to work herself out of a job.

Just some thoughts ... for whatever they may be worth.

Owen

On 8/25/2019 8:38 PM, Paola Di Maio wrote:
> apoogies this did not occur to me before - somewhat a bit scattered
>
> I d like to send out this late feedback to NIST, please feel free to 
> comment
> and make suggestions  (and add stratml version?) before the end of the 
> week
> Also, if anyone objects to the use of the plural We in the statement, 
> let me know
> and I ll use the singular I
>
> cheers
> PDM
>
>
> SENT VIA EMAIL
>
> To:  AI STANDARDS, NIST ai_standards@nist.gov 
> <mailto:ai_standards@nist.gov>
> cc:  Elham Tabassi, NIST elham.tabassi@nist.gov 
> <mailto:elham.tabassi@nist.gov>
>       and W3C AI KR CG
> Greetings NIST AI Standards-
>
> In relation to:
>
> A Plan for Federal Engagement in Developing Technical Standards and 
> Related Tools Prepared in response to Executive Order 13859 Submitted 
> on August 9, 2019
> https://www.nist.gov/sites/default/files/documents/2019/08/10/ai_standards_fedengagement_plan_9aug2019.pdf 
>
>
> First of all, thank you for this useful work. and apologies for 
> sending late feedback
> well past the June 10th deadline, I hope you can take it into account 
> anyway
>
> Given the definition on page 25
> / ISO/IEC 3WD 22989 Information Technology—Artificial 
> Intelligence—Artificial Intelligence Concepts and Terminology 
> artificial intelligence: capability of a system to acquire, process, 
> and apply knowledge /
>
> AI standards should include a socio technical standard for AI 
> Knowledge Representation
> Although still at exploratory stage, the W3C Community Group here is 
> currently gathering
> knowledge and resources towards that effect.
> https://www.w3.org/community/aikr/
>
> We respectfully request NIST gives due consideration for AI KR 
> standards, possibly open and web
> based.
>
> Best regards
> Paola Di Maio, Chair
>
>
>