Re: Interest in a zoom meeting on safe AGI?

Hi Dave,

I'm focused on Human Centric AI.  There's a process underway where a vote
is to be held soon in ISOC for an ISOC Human Centric AI SIG that i hope
will both come about; and then act to support the various 'social' aspects
that need to be worked through, as will also help to improve the W3C Human
Centric AI CG work.

With respect to the escalation of conflict internationally, you might also
want to consider what is needed to support the needs of people in relation
to IHL (international humanitarian law) and in-turn also, how designs may
influence whether or not and/or how, that may be made achievable.

Separately, i've started working on some considerations seeking to update
IHL related considerations to take into considerations the needs of people
now that we are so dependent upon technology.  Even digital identity to
protect refugees seems useless if when such tools are most needed, the
internet is turned off.

below are other notes.

kind regards,

Timothy Holborn.

It is my belief that there is no 'digital humanitarian' worker category
considered by existing IHL law and this is a significant problem that needs
to be addressed urgently.

It appears to me, that in-effect, IHL appears to be defined for a pre-ict
(internet/broadband / digital transformation) age and the requirements to
protect and promote peace infrastructure development, maintenance and
operation is at jeopardy, which leads to various significant threats..
As such, this is an unsolicited initiative of my own; and is work in
progress.  The documentation is still in its early stages, but I thought I
would seek to alert others ASAP as a matter that could be time-sensitive.

IF ANYONE IS AWARE OF EXISTING WORKS IN THIS AREA THAT ACT TO ADDRESS THESE
SORTS OF CONCERNS, I WOULD VERY MUCH WELCOME BEING ALERTED TO ANY SUCH
PRIOR WORKS.

In the belief that there is in-fact an absence of any such existing tools;
I have started to create some illustrative thoughts on potential
instruments;

https://docs.google.com/document/d/1215uqSsLbj3KACqXZIPSI-SQbg43AQ4VyixMK9K2InE/edit
<https://docs.google.com/document/d/1215uqSsLbj3KACqXZIPSI-SQbg43AQ4VyixMK9K2InE/edit>

That document has a few links in it, inc. whilst the initial document being
developed is from;

https://docs.google.com/document/d/1pl6UYIFmWFT1-AqOa7Bi2_eOrfwYBjRr0M7lg_YNl70/edit

NOTE: I have used Chat GPT-4 to instantiate some of the basic 'template'
works that then require significant development.

It is uncertain whether AI platforms like GPT-4 should be reasonably
classified a 'human rights' related 'humanitarian infrastructure'
requirement; given the potential for misuse of large-scale AI systems, both
those that are presently well-known and others either in development or
that may otherwise exist, but are secret.

*Thereafter; philosophically, much as computer operating systems have a
'safe mode' that reduces functionality to the basics in case of problems,
the same is seemingly required for humanity in case of problems also.  This
may well become part of the definition for 'human centric' ecosystems,
including but not limited to AI.   There are various implications related
to these considerations, including but not limited to those pertaining to
the availability and priority for the use of energy.*

Note that the document has a name and acronym for the initiative, that was
generated by GPT-4.

*Whilst, my first thought is that this looks like what i would call 'web
civics'; overall, i thought it would be best to leave the process of
figuring out which name best suits this initiative to a group process,
where i'd certainly be happy to submit both 'web civics' and indeed also
'trust factory' as options; but that, the intents & purposes are about the
functional requirements, not about 'putting my name on the building' or
similar, ideologically...  *

*There is presently an enormous amount of work to do; increased
international conflicts threaten my ability, and that of many other
parties, to continue to perform these works specifically for good
humanitarian purposes.   *

On Thu, 2 Nov 2023 at 01:07, Dave Raggett <dsr@w3.org> wrote:

> I would like to gauge the level of potential interest in scheduling a call
> on safe AGI.  This would be recorded and shared publicly after the event.
> Please let me know if you are interested in attending and what your time
> zone is.
>
> p.s. we could make this an open meeting for anyone to join.
>
> Many thanks,
>
> Dave Raggett <dsr@w3.org>
>
>
>
>