Transatlantic Cooperation on AI for Defense

Hodan, for whatever they may be worth, here are some comments on key 
points made by you and your panelists today 
<https://datainnovation.org/2021/03/how-to-deepen-transatlantic-cooperation-in-ai-for-defense/>:

  * Joanna - What does accountability mean in actual practice? She
    suggested that it means leadership.

Accountability requires records.  Scalable accountability requires 
machine-readable records. 
https://en.wikipedia.org/wiki/Machine-readable_document 
<https://en.wikipedia.org/wiki/Machine-readable_document>

  * Raluca - With respect to hierarchy, we need to think outside the box.

The traditional conceptualization of "leadership" reflects artificially 
ignorant 
<https://www.linkedin.com/pulse/artificial-ignorance-owen-ambur/> 
inside-the-box thinking.  It absolves us from taking personal 
responsibility for doing what we ourselves can do, in partnership with 
those who share our values and objectives. Brittle hierarchies and 
outmoded power structures <http://ambur.net/French&Raven.htm> should be 
supplanted by dynamic networks supported by online services enabling 
collaboration to realize common and complementary objectives.  How long 
The Politics Industry 
<https://www.linkedin.com/pulse/politics-industry-v-we-people-magic-formula-owen-ambur/> 
might be capable of thwarting the realization of such results-oriented 
networks (Communities of Results) remains to be seen.  Thus far, the 
federal bureaucracy here the U.S. has avoided accountability for 
compliance with section 10 
<https://www.linkedin.com/posts/owenambur_open-machine-readable-government-activity-6751906006801903616-9nt1> 
of the GPRA Modernization Act for a decade and counting.

  * Christie - The U.S. has published AI design principles and has one
    of the most transparent processes for the development of AI-enabled
    weapons.
  * Ulrike - The U.S. is actually developing more advanced AI policies
    that the EU is merely talking about, but Europeans are still
    skeptical of U.S. intentions.  There may be differences in values. 
    They should be reflected in strategies.
  * You - It is not clear what the EU strategy is.
  * Joanna - Can't really do much without an agreed upon plan and data
    is foundational.
  * Raluca - Referenced the relevance of international standards and the
    need for a vision of the future of defense as well as a strategy for
    AI within the broader context.
  * Christie - The U.S. is interested in working with allies who are
    interested in working with us.  Once a data sharing model has been
    developed, it can be rolled out to others.
  * Joanna - Begin by sharing non-sensitive data.

Strategy Markup Language (StratML) is the international data standard 
(ISO 17469-1 <https://www.iso.org/standard/59859.html>) for sharing the 
content of strategic plans, and StratML Part 2 (formerly ANSI/AIIM 
22:2017) specifies the additional elements required for performance 
plans and reports.  Such plans and reports are essential records for 
purposes of accountability.  The desire to move beyond talk to 
collaborative action is evidenced by the publication of performance 
plans with specified objectives, metrics, timelines, stakeholders, and 
interlinkages (enabled, leveraged, tracked and reported by value-added 
intermediaries).

Several of NATO's plans are available in StratML format at 
https://stratml.us/drybridge/index.htm#NATO

This Google site-specific query 
<https://www.google.com/search?q=artificial+intelligence+site%3Astratml.us&domains=stratml.us&sxsrf=ALeKk02nLqTGUiSml99j6D0-9K4stGKh5g%3A1616522979408&ei=4y5aYNqjGJeZwbkPmpi_6AI&oq=artificial+intelligence+site%3Astratml.us&gs_lcp=Cgdnd3Mtd2l6EAxQAFgAYPKwAmgAcAB4AIABmQGIAZkBkgEDMC4xmAEAqgEHZ3dzLXdpesABAQ&sclient=gws-wiz&ved=0ahUKEwiaodWvgcfvAhWXTDABHRrMDy0Q4dUDCA0> 
reveals that about 185 of the >5K files plans in the StratML collection 
reference "artificial intelligence," including the U.S. DoD's 2018 AI 
strategy <https://stratml.us/metaformixis/DODAIS.pdf>.

See also StratML use case Goal 30: Artificial Intelligence 
<https://stratml.us/carmel/iso/UC4SwStyle.xml#_6f069874-bb92-11e7-9b76-f79f9342c8d9> 
- Document on the Web in StratML format the performance plans of 
proposed artificial intelligence agents.

It will be interesting to see who truly wants to collaborate and how 
they plan to do so.  https://connectedcommunity.net/

Owen
https://www.linkedin.com/in/owenambur/