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.
With a tip of the hat to Abbie Hofmann and Steal This 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 skunkworks 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 Disruptive Technology. 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 the future.
Openness || Secrecy
As an OpenSource project like GC/MF (Google Code 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. Now the 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?
CounterEspionage
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 Sheeple and HoiPolloi who in blissful ignorance wallow in the distractions of MasPop entertainment.
Human Resources
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 ten years.
Suspicious Activity
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.
DeBriefing
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.
CyberSpace
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.
"so also the AI Mind must be kept from psychotic episodes and from breaking free of human control."
in other words, they are tortured.
great. just fucking great.
New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.
Keep up with the latest Advogato features by reading the Advogato status blog.
If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!
del.icio.us
Digg
Google bookmark
reddit
Simpy
StumbleUpon
Furl
Newsvine
Technorati
Tailrank