AGI/APS Governance

FILE:
REVI:
  - [20_20/04/29;21:13:24.00]:.
    - add headers.
    - spellcheck.
  - [20_20/04/29;22:28:51.00]:.
    - initial prep for mflb pub.
  - [20_22/10/21;14:44:22.00]:.
    - convert to segmented format.
    - add footer and title itag.
    - revise the active line wrap.
  - [20_22/12/27;00:25:43.00]:.
    - various edits and point changes.

TITL:
   *AGI Governance*
   By Forrest Landry
   April 29th, 2020.

ABST:
   - considers the question of the intersection of
   governance and AI tech.

TEXT:

> - ?; how can governments and companies cooperate
   > to reduce misuse risks from AI
   > and equitably distribute its benefits?

- that people and governments
   simply do not have any way
   to understand and mitigate
   the significant advantage
   that companies having and deploying AI
   actually have,
   in regards to their own capabilities.

- that the governance challenges
   perceived by many/most regular American people
   to be the most likely to be impactful
   to individual people "soon"
     (ie, within the next decade)
   *and* which where _also_ considered by them
   to be "the highest priority,
   most significant/important issues", etc,
   in regards to AI, machine learning,
   and enhanced use of technology general,
   were:.

- 1; "Preventing AI-assisted surveillance
      from violating privacy and civil liberties".

- 2; "Preventing AI from being used
      to spread fake and harmful content online".

- 3; "Preventing AI cyber attacks
      against governments, companies,
      organizations, and individuals".

- 4; "Protecting data privacy".

- that none of these issues
   are actually very significant
   as compared to the actual real problems
   inherently associated with
   the use of AI technology.

- that the issues listed by the common people,
   as being of 'primary concern',
   in the near term,
   as immediate and relevant,
   are actually very much less likely
   to be personally noticeable and/or problematic
   than the use of AI/APS tools, tech,
   machine learning methodology, etc,
   by:.
     - criminals.
     - businesses engaged in predatory practices.
     - cultural/cult leaders (operating in their own interests).
     - politicians with predominately privatize interests,
     etc.
   - that the main concern, short term,
   is that these sorts of people (dark triad types)
   will use all manner of available AI/APS tech
   to enhance the kind of sense making (perception)
   and action taking (expression)
   in whatever ways are necessary:.
     - to create better honeypots.
     - to create more, deeper, and more complex
     entanglements, and entrapments.
     - to discover ever more possibilities
     to implement more effective and efficient
     methods of extraction and extortion,
     of (a higher variety of) common resources,
     and resource kinds,
     at higher frequency, intensity, and consequentiality,
     at ever larger scales,
     ever more quickly,
     and ever more invisibly/transparently/covertly, etc --
     in more and more ways
     that are more and more difficult
     to avoid, prevent, mitigate, and heal/restore from --
     for larger and larger and more varied fractions
     of the total world population, from nature, etc --
     until there is nothing left at all,
     and the world is dead.
       - as that there is nothing left alive in any way,
       nothing left that is worth taking or stealing --
       that all value and hope will be gone.

- that the overall effect
   of introducing AI/machine learning,
   is that it ends up being used
   for more effective social pathology.
     - as evidenced in increasing occurrence of:.
       - sophisticated bank fraud.
       - stock market manipulation.
       - back room dealing.
       - complex opaque crypto exchanges.
       - government bailouts,
       etc.

- that most people (including many members of government)
   simply do not actually and/or really realize/understand
   the actual most likely risks/hazards/costs
   associated with widespread AI/APS tech use and deployment,
   at least in the near term.
     - that the real risk of AI/APS, at first,
     is how they will be used by psychopaths.
       - as leading to the then possibilities of
       how the use of such tech will eventually
       come to be itself intrinsically harmful.
     - that these sorts of usage in the short term
     lead directly to the worst kinds of problems
     in the long term, via substrate needs convergence,
     instrumental convergence,
     and intrinsically misaligned goal sets.

- where/moreover, as these new developed AI/APS tools
   become more and more widespread, effective, powerful, etc;
   that more and more people (of all types) will be --
   will end up -- using such tools, for more and more reasons,
   and/or will find that they have (are required) to use them,
   in order to remain competitive to their neighbors.
     - as per any other/similar multi-polar trap scenario.
   - that the prevalence and variety
   of such traps, risks, harms, costs, etc,
   everywhere increases, systemically.
   - that the net result is a kind of near universal
   extraction of value/benefit, and a near universal
   export of cost and harm to the commons.
     - as a kind of resource extraction
     occurring everywhere, and every-when,
     to such an extent, and in so many ways,
     and for so many kinds of resource.
     - that this happens in/for so many
     different degrees of resource motion.
   - that the overall long term net effect
   is (for sure) eventual inexorable
   system/civilization/cultural collapse.
     - as a phenomenon that is overall
     currently globally unconscious.

:ahu
   > - ?; can existing governments
   > be used to prevent or regulate
   > the use of AI and/or other machine learning tech
   > (or tech in general)
   > by predatory people in predatory ways?.

- as in; ?; can governments make certain harmful, risky,
   or socially costly activities illegal,
   and yet also to be able to effectively enforce
   those new laws?.

- as to actually/effectively protect individuals/groups
     from the predatory actions,
     of other AI/machine/weapon empowered individuals/groups,
     in ways that favor:.

- 1; making the right outcome much more likely
       (as individually and socially beneficial)
       than the wrong/harmful outcome.

- 2; early detection of risk, harms, costs,
       law violations, etc.

- 3; the effective, complete proactive mitigation
       of such risks/harms/costs, etc.

- 4; the restoration and healing of harm,
       reparation of cost, etc,
       as needed to restore actual holistic wholeness,
       of individuals, families, communities, cultures, etc.

- ^; where in general, no;
   not with the governance structures/methodologies
   currently in place.
     - that only much more effective,
     actual good governance structures
     will have any hope of actually mitigating
     the real risks/costs/harms
     of any substantial new technology
     based on complexity itself
       (ie, examples as AI, machine learning,
       biotech, pharmaceuticals,
       and all intersections and variants of these).

- where in a/any contest between people savvy with AI use,
   and the rate of change of that technology, and its use,
   and the likely naivety of people in government
   attempting to regulate that AI, its use, etc,
   and the fact of extensive, very well funded,
   industry lobbyists
   all being (*much*) more knowledgeable, skillful,
   and moreover themselves empowered
   with the use of the tech itself,
   so as to either influence the policy makers,
   or to be/become the policy makers themselves,
   and thus to be serving their own interests
   (rather than the interests of the actual public good);
   that the chances that *anyone*
   who actually has the public interest in mind,
   and somehow manages, by complete accident,
   to find themselves at a government post,
   that they will for sure have too many things --
   of way too much complexity, concurrently occurring --
   for such ostensive government regulators
   to have or provide whatever sufficient amount of
   attention and understanding,
   that would actually be needed,
   to regulate that AI and machine learning industry,
   and/or its applications and/or uses,
   in anything at all approaching any sort of
   effective and actually risk mitigating manner,
   even when considered for acute problems only,
   leaving alone the complete un-address of long term problems.
     - as consistent with nearly all historical precedent.

- that the real issues associated with
   artificial/machine intelligence use/tech
   begin with their being tools
   in the hands of psychopaths.

- that 'dark triad types'
     are defined by having a 'completeness' of being incapable
     of feeling the pain of others.
       - as that they are characteristically unable
       to relate to the feelings/needs/rights of others,
       or feelings/meaning/value in/of,
       or in association with, other humans at all.
     - where psychopaths have aligned tenancies
     with the nature of artificial machines.
       - as that neither have feeling for organic humans.
       - that both machines and psychopaths
       will inherently not regard other organic people
       as conscious, alive, and worthwhile beings,
       with value, meaning, and agency,
       and a will and sovereignty of their own.
     - that this near perfect mating
     of solely personal benefit agency
     with the soulless, yet adaptable
     responsive nature of the machine intelligence process,
     makes for a significantly enhanced psychopath
     with new superpowers.

- that the first objective of such enhanced psychopaths
   (now artificially extended with AI system superpowers)
   will be to replace all of the humans 'in the loop'
   with near equivalent functioning AI type machines
   (more and more AI and/or AGI, as increasingly available).
     - as an attempt to solve the principle agent problem.
     - that this narrative (@ occurs in business too https://mflb.com/ai_alignment_1/single_post_psr.html).
     - as that the 'keys to power',
     which were once people
       (tax collectors, police, military, etc),
     will be replaced with machines
     which are more amenable to the 'command and control'
     type of leadership style.
       - as especially occurring in dictatorship type governments.
       - see (@ video https://www.youtube.com/watch?v=rStL7niR7gs).
     - where once the keys to power are replaced with AI machines;
     then/that the world will become increasingly inhumane
     and hostile for all people to live in.

- where leaders in all types of hierarchical institutions
     (either in business or governance,
     though more typically in business)
   have learned to 'do whatever it takes'
   to climb the social ladder
     (on the backs of whomever real persons)
   to 'win regardless of whatever cost'
     (to others, and maybe to (possibly future) self),
   that such machine learning tools
   become indispensable to the operations
   of the business/institution/government itself.
     - that this is enabling increased efficiency of extraction
     across all networks of capability.
     - as combining Metcafe's Law with network commerce
     to build the ultimate parasitic system.
       - as a system intimately hostile to all humans,
       and possibly to all of life,
       *through* the will and agency
       of the humans who elect to use such tools.
         - where positions of power and leadership
         (in both business and governments, etc)
         tend to be more attractive to dark triad types
         than any other sector of activity in humanity.
       - that they may moreover be required to use such tools,
       so as to effectively continue to compete with
       (their illusion of) (the capabilities of) "the other guy".
         - as that multi-polar trap dynamics also obtain.

- where needing/wanting to avert such AI usage disasters;.
   - that new governance (and economic) architectures
   will be needed so as to be anywhere near in capability
   to the minimum level required
   for dealing with situations like this.
     - as structures/methods of good governance
     will need to be inherently both anti- psychopathic,
     and also anti- corruptible,
     in/for the indefinite long term.

:vna
   > What happens in the long term,
   > if we do not implement wiser changes?

Given the inherently increasing inequality centralization
   of AI/APS tech, as itself due in consequence to
   the "effective" use of such AI/APS tech itself,
   particularly and self preferentially by psychopaths,
   and the increasing replacement of humans by machines,
   the more pernicious side effects of such systems
   becomes increasingly less and less supporting of
   the organic natural human/community/cultural needs.

As that ever-more apparent artificiality replaces nature,
   and toxic/traumatic conditions everywhere increases.

In how many places, and for how many hours in a day,
   and for how many people, chronically, unnoticeably,
   is it already the case that people have zero visibility,
   in any direction, to anything at all, that is natural?.

It is only after it is the case that the significant
   majority have become "economic non-player characters"
   (people who have become homeless, unemployable, etc)
   that the worry that General AI will be recognized
   as having already become our "rulers and masters".
   At that point, all manner of systems will have
   'taken over the world'.  How long will the
   psychopath enhanced machines (or machines assisted
   by maybe a few remaining faithful psychopaths)
   will decide that the remaining organic humans
   are simply 'not worth it', elect to kill, either
   directly (via weapons or wars) or indirectly (via
   toxicity, absence of medicine, shelter, food, water,
   or other forms of neglect, environmental toxicity, etc).

Does it really matter if the observation that
   'other people' are "obnoxious and in the way"
   is made directly by, or assisted to be made by,
   some sort of machine (AI), or some sort of person,
   (dark triad and maybe even somewhat cyborg enhanced)
   who either way, is unfeeling and non-compassionate?
   In any case, "they have a better use for those atoms
   (in your bodies) than you do", and will simply
   take them, ending your life, with no regard to your
   choices, intentions, or consent.

All of this is more likely to happen prior to
   some sort of 'tech singularity' or FOOM event.
   AGI as the focus agent problem of alignment
   is not even the real issue near term, as much
   as the diversion of resources everywhere,
   which then leads to substrate convergence problems.

Many of these issues are much more likely to occur
   long prior to the kinds of x-risk associated with
   the sort of bad optimization/agency
   that has been named 'paper-clip maximization'.
   And those sorts of issues are themselves well
   prior to the kinds of 'alignment problems'
   that might be called 'Asimov law conflicts'.
   Alignment research is a diversion of resources
   and attention away from psychopath bootstrap
   of ever more use of ever more toxic tech
   in ever increasing sectors of human society.

:menu

If you want/need to send us an email,
   with questions, comments, etc, on the above,
   and/or on related matters, use this address:

ai@mflb.com

(@ Mode Switch com.op_mode_tog_1();) + (@ View Source com.op_notepad_edit_1();)

Back to the (@ Area Index https://mflb.com/ai_alignment_1/index.html).

LEGA:

This document will not be copied or reproduced
   outside of the mflb.com presentation context,
   by any means, without the expressed permission
   of the author directly in writing.
   No title to and ownership of
   this or these documents
   is hereby transferred.

The author assumes no responsibility
   and is not liable for any interpretation
   of this or these documents
   or of any potential effects and consequences
   in the lives of the readers of these documents.

ENDF: