Re: AI & Democracy

On Fri, Apr 26, 2019, 7:04 AM Paola Di Maio <paola.dimaio@gmail.com> wrote:

> Thank you Owen
>
> My view is that humanity has not yet fully matured/emancipated, ie has not
> evolved
> to dominate their mind and control emotions and not to lie (too much) etc
>

I Agree.

Once upon a time, in one of this lists, someone asked how one may state "I
am" in terms of a semantic web statement. My reply was:

"iAm, complementOf, everythingWhatImNot".

The inverse could be stated by swapping subject and object.

I think the meaning of this is being able to recognize unity and disparity,
and knowing the boundaries between them. With 'boundaries' I means 'rights'
and 'obligations' between pairs of united / separated individuals (by their
preferences, for example) being 'preferences' the means that orchestrate
our interests and interactions.


> Political institutions reflect the state human condition reflecting
> strengths (aspirations, good intentions) and limitations (....fallibility
> etc.list....)
>

Human political institutions are composed by humans. I agree.


> Democracy would be nice but I am not sure we have found a way to
> materialise it as such
> I believe that technology can help a lot,
>

Maybe, I don't remember its exact name, but a balloting system where voters
choose their candidates in order of preference (i.e.: not choosing one
instead of another but 'sorting' them by their preferences) could help to
alleviate the I-Voting and the "mayority" problem.


> Now we need to look at what we mean by AI
>

AI should be to Human Intelligence like a bicycle is to human legs.


> I don''t agree with what most people call AI at all,  for example I am
> struggling to explain to a friend that I believe AI  should extend and
> improve human capabilities, not replace them or attempt to reproduce human
> features to replace humans (that would be a disaster)
>

AI was / is developed by HI, obviously reflecting HI intentions and to
enhance its capabilities. Disaster (sci-fi like) would be AI having it's
own intentions and developing further intelligence...


> If we can develop an AI that can help humans to become better (*more
> evolved, better informed, less afraid of telling the truth etc) then I am
> sure AI can also help
> with governance
>

The problem (AI + Big Data of everything) is who controls or have access to
this overwhelming stream of indicators. Or when decisions are delegated to
machines because of this enormous amount of data. We humans, actively or
passively, are not prepared for the truth about everything every time.


> A lot of work to be done there, and that is why I take an interest in the
> subject
>

We have very evolved Data systems (passive facts).

We are currently developing and using advanced Information management
systems (Business Intelligence and Big Data inferences).

We are scratching the surface of Knowledge aware systems, those who can
reason and infer like human does.

Once this layers stack will be mainstream and fully functional, care should
be taken. We could implement 'crowd-prompting' in Decision Support Systems
for, for example, sensible (ethics) decisions feedback.

Best,
Sebastián.



> Lets us have a link to your article
> :-)
>
> PDM
>
>
>
> On Fri, Apr 26, 2019 at 12:00 PM <owen.ambur@verizon.net> wrote:
>
>> This article suggests AI may threaten democracy:
>> https://gcn.com/articles/2019/04/24/ai-threats-democracy.aspx?s=gcntech_250419
>>
>>
>>
>> However, to the extent that democracy is majoritarianism (rule of the
>> majority <https://en.wikipedia.org/wiki/Democracy>), AI might be used to
>> increase its effectiveness – both in shaping the views of the majority
>> (groupthink) as well as oppressing minorities who fail to adopt the
>> majoritarian point of view.  A case might be made that not only the social
>> media but also the so-called “mainstream” media have been doing so.
>> Conservatives certainly believe that to be the case, with more than a
>> little justification in light of the well-documented biases of the media..
>>
>>
>>
>> Thus, democracy itself, unfettered, might be a  bigger part of the
>> problem than the solution, and while Churchill may have been right in his
>> time, he did not live in the cyberage.
>> https://richardlangworth.com/worst-form-of-government
>>
>>
>>
>> I’m thinking about writing an article entitled “Beyond Democracy and the
>> Rule of Law” positing that we:
>>
>>
>>
>> a) now have the means to do better than majoritarianism, and
>>
>> b) already have too many laws, regulations, and guidelines imposed from
>> the top down in narrative format and far too few actual self-expressed
>> performance plans published in open, standard, machine-readable format to
>> be socially supported and enforced by peers.
>>
>>
>>
>> At a minimum, the tech giants who are helping authoritarian governments
>> adopt AI should be expected to develop and openly publish algorithms
>> addressing the rules under which they believe it may be appropriate for
>> anyone to attempt to impose their will upon others.
>>
>>
>>
>> See also http://connectedcommunity.net/ & http://aboutthem.info/
>>
>>
>>
>> Owen
>>
>