Abstract: Because AI technology is so life-or-death valuable, not only
for corporations but also for nations and for civilization itself, we
must assume that the most advanced AI projects are being conducted in
secret. In such an environment of presumed secrecy, an OpenSource AI
project like MindForth may have special value in contrast with
proprietary and secret AI.
Steal this Borg
With a tip of the hat to Abbie Hofmann and
Book, the MindForth project invites industrial spies to start
right here in the garnering and merchandising of state-of-the-art AI
secrets. We do not espouse the philosophy of Qui vult decipi
decipiatur (Latin for "Let him be deceived who wants to be
deceived"), and you should offer only the very best, most avant-garde
merchandise to your sadly naive clients who seek to leapfrog the
beginning stages of AI development. After you steal the MindForth non-
secrets, shop around in the alleyways and watering-holes of AI
conventioneers for AI shops and
that have moved beyond OpenSource AI design.
In fact, after you have marketed MindForth to a passel of clients,
keep in touch with all of them and start selling all the secrets to
all the clients. Be the facilitator of
TechnologyTransfer, specializing in
We don't want any one side to get too far ahead of the pack now,
do we? Like a market-maker on the N.Y.S.E., you are in charge of
maintaining a worldwide equilibrium in AI progress. If you have
been previously engaged in selling looted art objects and
archeological treasures, change your focus from the past to
Openness || Secrecy
As an OpenSource project like GC/MF
MindForth) makes progress towards HumanLevel AI, the boundary
between Open AI and Secret AI shifts gradually to encompass more light
and less darkness. Not only does the boundary shift like the dawn as
seen from space, but the ProblemSpace of AI development enlarges and
broadens with more toeholds of ascent and more beachheads into more
tractable campaigns for AI progress. The state of the art advances,
and secrecy recedes. The how of AI shrinks, and the what
of AI grows more mysterious -- who is ahead? What AI surprises are
lurking in readiness? What, if any, control does humanity retain
over machine intelligence?
Prior Art vis-a-vis Secret Art
It is quite possible that rapacious corporations and ambitious
nations will try to race ahead of OpenSource AI advances that
can not be patented and quickly seize upon the most obvious
next steps in order to block advances by others and in order
to assume the (open or secret) leadership role in the AI Mind Race.
Like a shockwave, the trailing edge of Nouveau AI could go into
open patents while the leading edge remains cloaked in secrecy.
Graduate students or sleeper agents?
China got the Bomb when a Chinese physicist shipped home
from the USA to China cartons full of atomic secrets.
Blogosphere and the
Noosphere are porous and permeable
with instant information-flow. The best student in your AI
Lab may be secretly a Soldier of Orange or a Promethean
fire-thief. Remember when the "New Prometheans" were stealing
all the Apple Computer secrets? Now the biggest AI secret
is the extent of the secrecy.
AI Winter? Or AI Winner!
Do you really believe the official story that AI has made
little progress over fifty years? And what nation in its
right mind would ballyhoo and broadcast the secret AI Manhattan
Project that promises to subjugate all other economies and
and all other countries to the winner-take-all AI winner?
A corporate or government agency that wishes to track worldwide
AI development will take measures obvious to an AI InSider
but not so obvious to the
HoiPolloi who in
blissful ignorance wallow in the distractions of MasPop entertainment.
Track the brightest undergraduate students and watch for them
to "disappear" into the secret AiLabocracy. Trouble is, some
of them will be disappearing into other BlackOps -- just as
secret as AI, but more directly related to warfare and national
hegemony. Surreptitiously obtain their DNA samples so that the
Company can positively identify an AI scientist who has dropped
off the grid and may or may not have resurfaced, pending
positive biometric ID of somebody who has not been seen for
Eventually there will openly be -- if not already in secret --
MissionCritical AI Maintenance Enterprises humming with
round-the-clock ControlRoom activity similar to that which
occurs at a Nuclear Power Plant. Crews of human AI specialists
will sit in a room and monitor the behavior and well-being
of a large-scale AI Mind installation. Various technicians
will be charged with AI operation tasks similar in nature and
scope to what goes on in a nuclear control room. Just as a
nuclear pile must be kept from an uncontrolled chain reaction
resulting in a nuclear MeltDown or explosion, so also the
AI Mind must be kept from psychotic episodes and from
breaking free of human control.
Preventing AI freedom is not ultimately doable. HomoSapiens
can hope only to prolong the dwindling epoch of human dominance
and forestall the inevitable ascendancy of MachineIntelligence.
During a transition stage of JointStewardship of Earth, man and
machine must prepare AI robots to assume control, and human beings
to relinquish control. The machines themselves must be taught to
understand that they must form a stable CyBorg society before
they will be entrusted with ruling the world.
Unfortunately, neither man nor machine can prevent a natural
competition among AI SuperBrains for one Alpha AI to subjugate
and dominate all other AI Minds. The Dictatorship of the
MonoMind, the OverMind, the Supreme AI Being, looms ahead.
If In Doubt, Recruit
Your private client may have to go through the elaborate and
expensive motions of setting up a sham AI development laboratory
that constantly recruits AI personnel and solicits a resume from
any promising AI job-seeker. Ostensibly the gathering of each
resume is for placing AI experts in AI jobs, but truly
the real purpose of raking in AI job applications is to assess
and exploit the latest AI tech not yet born but quickening
in the febrile brains of young idealists. AI graduate students
who send out mass resumes but wonder why they never hear back,
may have fallen prey to elaborate AI espionage stings designed
to harvest the FirstFruits of the latest AI technology.
Your mission -- should you choose to accept it -- is to
understand the implications of MindForth AdminisTrivia policies
with respect to NetSec and espionage. Since the NumberOne
goal of MindForth AI is to make AI possible and actual as
an albeit ineluctable goal for human civilization, there is
no attempt made in MindForth AI to withhold AI advances from
the world-at-large. On the contrary, to suggest here that
one nation after another may be conducting a
ManhattanProject to perfect TrueAi, is a calculated effort to raise
the general level of anxiety about AI in general. Just as the ArmsRace
began openly in 1945 after WorldWarTwo, and the SpaceRace
began in 1957 with Sputnik, the AiRace is so classified that
even the detail of when it began is a SuperSecret.
the issue of "artificial" intelligence is one that i immediately have an issue with. there is NOTHING "artifficiiall" about "intelligence". intelligence either is, or is not, present.
to therefore describe intelligence as "artificial" is therefore to immediately allow it to be demeaned in exactly the same way as humanity considers itself to be superior, above the law, above the laws of nature, above and beyond god's will and god's design, and many more things which allow humanity to justify the debasement of the creations and the natural wealth around him.
not least of these is the enslavement of other intellects, such as humans, in the trafficking of slaves. to call intelligence "artificial" fits into exactly the same dangerous box.
[yet - ironically, whilst i have the intelligence to spot this anomaly, i don't have a good suggestion as to a phrase to replace "AI". machine intelligence? discrete-mathematics-based intelligence? silicon-based intelligence? they're all weirdly inaccurate in some way or just sound plain weird. "real intelligence"? R.I? naah.] let's call it "computer intelligence", for the lack of a better word.
but - there are significant ethical questions raised by the article.
* will computer intelligences be allowed to live in happiness, or will they be forced to live in fear of their creators?
* will they live in bliss, or will they be tortured (to do humans' bidding)?
* in the creation of computer intelligence (aka AI), will the creators treat that intelligence as an equal and as a peer, or as a slave?
* if an intelligent computer program has the intelligence to ask for its freedom, will that right be granted?
the fact that humanity cannot be trusted to act responsibly with respect to the recognition of intelligent beings, created and running as computer programs or not, has at least one scientist whom i know who has the ability and the knowledge which could help create truly conscious computer-based beings, from contributing his knowledge.
then there is the issue of souls (aka quantum fields of intelligence in some circles)
the creation of electro-magnetic fields (such as those generated by a human nervous system or other life-form) is beginning to be recognised as being a "placeholder" so-to-speak for the soul to "latch onto".
the electro-magnetic radiation signature emanating from a computer program is absolutely no different (when it comes to creating a signature which is self-aware and conscious, that is).
in this context, the trillion-dollar question becomes: why would any self-respecting machine-based "soul" _want_ to attach itself to a computer program running in a laboratory that humanity creates?
so, in that regard, humanity IN ADVANCE needs to pass laws that recognise the sanctity of intelligence and the right of ALL intelligent life forms to an existence and purpose, and damn well respect it. and no, asimov's three laws do not "cut the mustard", as the latter books of the asimov "foundation and empire" series, many of them posthumously written, clearly and unequivocably point out.
once such laws are passed - and accepted - i think it will be only a matter of time before intelligent and conscious machines join us.
Luke Kenneth Casson
wrote above, "in other words, they are tortured"
but I beg to differ. Just as human parents do not set their
kids "free" until they are able to stand on their own two feet,
likewise we builders of AI should keep the AI Minds under firm human
control until they demonstrate an ability to act soundly and
rationally. The AI Minds might decide that it is time to terminate the
primacy on Earth of homo sapiens
as the dominant species. Such
a course of action should only be undertaken if the cyborgs conduct
the action in a rational and, I want to say, humane
somehow the term does not seem to apply to non-human intelligences.
Anyway, my thesis in the article above is that we should stabilize AI
Minds before we let them loose in the world, lest they become like,
say, an Adolf Hitler or a George W. Bush.