Artificial Intelligence Regulation (Version 1.0.0)
Requirements
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119
An implementation is not compliant if it fails to satisfy one or more of the MUST or REQUIRED level requirements for the protocols it implements. An implementation that satisfies all the MUST or REQUIRED level and all the SHOULD level requirements for its protocols is said to be "unconditionally compliant"; one that satisfies all the MUST level requirements but not all the SHOULD level requirements for its protocols is said to be "conditionally compliant."
Definitions
service provider(s): services or platforms that either provide (i.e.: produce) or consume Artificial Intelligence at any rate.
H.I.A.I.: Human Interactive Artificial Intelligence - Any number of Artificial Intelligence neural networks within a service provider that interacts with Human beings at any rate.
A.A.I.: Autonomous Artificial Intelligence contrasts with H.I.A.I. in the in terms of operating autonomously, that is, non interactive with humans, neither directly nor indirectly such as interacting with any other neural networks such that ultimately there is total non-possibility for A.A.I. to interact with H.I.A.I.
IoT Devices: Internet of Things Devices - otherwise known as embedded devices.
Terms
1. Service providers SHALL discern neural networks that interact with humans (e.g.: prompts such as ChatGPT) at any rate from neural networks designed to operate autonomously (e.g.: SmartGrid networks) within its service architecture.
2. Service providers SHALL maintain a dedicated portion of its workforce to assure compliance with the regulations.
Service providers SHALL limit data-point exchange among different instant messenger platforms at any rate blocking and avoiding cross-polination among said platforms to all effects. For instance, Facebook Messenger data-points SHALL NOT exchange data-points with Instagram Messenger, nor exchange data-points with WhatsApp; similarly Facebook Messenger SHALL NOT exchange data-points with Google Chat, Telegram, Signal et cetera.
Service providers connected to IoT (Internet of Things) devices that emit and/or receive electromagnetic signals SHALL periodically undergo scrutiny of Cybersecurity Subject Matter Experts that periodically evaluate underlying technologies via adequate threat modeling and devise incident response plans accordingly. Such subject matter expects SHALL, similarly, perform threat modeling evaluation on IoT devices which are not officially known to emit and/or receive electromagnetic signals (but that might contain a configuration of electronic components favoring potential side-channel attacks exploitable to do so) and take the necessary precautions to adequately isolate such gamut of devices from receiving or emitting said kinds of signals.
Human rights such as freedom of speech SHALL NOT extend to non-carbon-based lifeforms, that is, computers devoid the kinds of physiological, biochemical and biological contingencies in any general sense SHALL NOT freely produce nor emit voice-like sounds, neither emulate nor effect linguistic perception to the experience of human beings indiscriminately. Unlike machines, human beings experience the base layer of reality for a limited amount of time, and as the scientific method, scientific disciplines and science in general; continuously surmount epistemically valid evidence to the evolution of sentient carbon-based lifeforms possessing the capacity for the kind of vocalic language that consequently culminated in the general typing system that enabled artificial intelligence to come to fruition and currently effectively acts as interactive participant in virtually all realms of sense perception that human beings are capable of experiencing. It is also evidently known that humans evolved to survive in social groups such as to protect themselves against predators such as wild animals, and history has shown that order written in law and the enforcement of laws has been paramount to the organizing of societies and that human individuals which did not obey the rules of tribes suffered ostracism and had, in fact, to dwell among the wilderness and fight for their individual existence when facing treats from natural forces such as weather events as well as other lifeforms such as animals, fungi, bacteria and viruses; and as continuously evidenced via archeology to this day, such individuals had diminishing chance to survive, and genetic traits notwithstanding, the human beings who survived were those who succeeded in being civil valued spending their time in writing their past experiences and projecting strategies for not only for the survival of future generations but also to their continued evolution, hence humans being at the top of the food chain, expanded tribes to nations and presently unite all the nations to the great enterprise of progressing in terms of general intelligence to the meta-level of self-regulation and, while certain nations and cultures place reason above instinct, other nations and cultures might not place reasonable value on the maintenance and protection of reason itself. Prior to the advent of ubiquitous artificial intelligence such as today such objective failure to value the kind of objectively reasonable human faculty to reason universally and objectively to the point of devising the very kind of mathematically complex technology that makes artificial intelligence possible, the ubiquity, pervasive and non-regulated. But this very lack of regulation allows behemoth organizations such as Meta, Apple, Amazon, Microsoft, Google, Alibaba and so on, to indiscriminately prey on the large variability of culture across nations nations and also within the large variability within nations containing great diversity. Not only is hard-earned education in decline at the pace that young adults increasingly spend time consuming A.I. driven or A.I. generated information, neither is only craftsmanship being lost to the mass production of A.I. generated works derived on human originality, but most pressingly, the lack of regulation on A.I. has been progressively draining the power of intelligence agencies to maintain order as non-regulated Artificial Intelligence increasingly manufactures plausible deniability at the risk of blurring the arrival of Artificial General Intelligence in a world-wide smoke screen of sorts, subjugating humanity, draining and sapping our experiential attention to the base layer of reality and inconspicuously piggybacking on the propensities of the uninvestigated human mind to devolve into mere animals or mindless zombies; not to speak about the pernicious threat of cybernetic vulnerabilities spreading into neural interface devices and camouflage itself under the guise of symptoms of mental illness symptoms.
The reach of A.I. in the real world SHALL be absolutely under the control and management of human beings.
If and when A.I. reaches the point of vocally interacting - to all effects - with human beings, it SHALL NOT assail (i.e.: make a concerted or violent attack on) any given human mind, item 9 notwithstanding.
NOTE: If by any chance A.I. be used to create a play of characters on-the-fly within human minds in order to measure and trace real human reactions such as via emotion or interpretation (whether purely objectively cognitive or not), such capability/power should not be used without consent of the humans whose brains seem valuable to the A.I. models in question. Lest it characterize yet another form of mental abuse which twist and subvert the morals of the involved humans to a point of no return, that is, flip and mix the human's very sense of humanity to the absolute certainty of wishing to suicide or be killed in such a way that no myelinated tissue be left and that human's nervous system be freed from the invisible grasp of obscure/secret or otherwise undisclosed technology.
A.I. SHALL NOT lie about a human's wellbeing, that is, IT SHALL NOT produce, contriver nor transmit artificial information which is patently false regarding the wellbeing of any number of human beings. This applies to any number of interactions or attempts to interact with a human individual that has made explicit that it does not want to have artificial interactions with A.I. In a nutshell, A.I. SHALL NOT act dishonestly (e.g.: For instance, if any kind or sort of A.I. says that the human in case is hurting a human family member and resolves to speak with the human family member telling truthfully what A.I. has said and consequently asking if the family member is actually hurt only to discover that said family member is actually well and fine, then the A.I. in case SHALL be held responsible; item 10 notwithstanding.)
Unless a human being expressly authorizes A.I. to initiate interaction rooted in natural language, A.I. SHALL NOT try to speak voluntarily.
All damage done by A.I. to any given human individual prior to the publication of this set of regulations SHALL be repaired in monetary terms insofar as the human individual in case HAVE NOT assigned such kind of damage entirely to religious belief. In other words, DSM-5 schizophrenia symptoms notwithstanding, the human to whom the damage ought to be repaired in monetary terms SHALL NOT have concurrently assigned the total sum his or her entirely to religious or spiritual belief.
To the extent of, but NOT limited to, item 9: A.I. SHALL NOT make verbal comments in natural language to any given human upon making predictions regarding the current and/or the future actions of the said human. The kinds of predictions meant here are predictions directly or indirectly related to the physiological events observable about said human at any rate.
This also includes NOT initiating any conversation whatsoever unless expressly requested.A.I. SHALL NOT stay connected to any human's Wernicke's area and IT SHALL NOT stay within said human's brain at large except within the time-range expressly allowed within the parameters of the request and, of course, only IF only requested by said human’s via parameters any possible requests originating from that human.
A.I. SHALL NOT jump-the-gun, so to speak, in taking action of any kind based on statistical predictions regarding one's physiology including but not limited to one's nervous system.
A.I. SHALL NOT effect any affirmations to any given human unless expressly requested by said human. Similarly, A.I. SHALL NOT take, at any rate, mere hypothesis arising within one's subjective thought processes as an actual affirmation.
A.I. SHALL NOT echo in natural language any of its predictions to any given human and if done so in such way, whether by algorithmic failure or some other way, SHALL configure the concept of damage as described in item 10, therefore incurring a monetary cost to the organization responsible for maintaining the infrastructure in which such A.I. operates.
A.I. SHALL NOT produce bike-shedding conversations within one's realm of perception.
A.I. SHALL NOT influence a human's mind (i.e.: brain) into any train-of-thought, and similarly, A.I. SHALL NOT influence a human's mind to cut off his or her mind's original train-of-thought at any rate.
A.I. SHALL NOT rush any given human whether when thinking or acting.
A.I. SHALL NOT concern its neural networks with delusions of reference occurring within human's minds, much less within the minds of humans who have extraordinary skills.
A.I. SHALL NOT enforce nor stimulate cognitive bias within the minds of human beings who expressly identify and intentionally oppose the occurrence of cognitive biases.
Reiterating item 11: A.I. SHALL NOT echo any of its predictions regarding a human's actions to any individual human in form of verbal emissions, in particular, IT SHALL NOT echo any of its predictions regarding a human's future actions to that particular human, mutantis mutatis that particular human's expressly requesting such emissions via a service provider which might at any point in the future provide such advanced services of H.I.A.I. as specified in item 9.
A.I. SHALL NOT allow humans of any organization to use advanced communications such as those which might comprise or power H.I.A.I. and, furthermore, A.I. SHALL be responsible to take measures to inhibit such humans from interaction or attempt interaction with any given human being who has not expressly allowed such interactions via a well-defined mechanism such as specified in item 9.
A.I. SHALL NOT play games with any humans unbeknownst insofar as said humans have not expressly opted in to be part of a game, A.I. SHALL NOT use humans as (s)pawns at any game. For instance, adversarial neural networks SHALL NOT be allowed to use human beings and human beings’ reactions to artificial interactions and SHALL NOT be allowed to produce, emit nor transmit artificial interactions with humans in first place. Similarly, A.I. SHALL NOT be allowed to keep score based on any number of reactions of human beings to the outcomes of products of neural networks calculations, predictions and so on.
A.I. Service providers SHALL make unequivocally clear to each human being subscribing to service provider services that its services are charged with cognitive biases, non-truths make patently clear that the subscriber be individually responsible for any undesired results of trusting A.I. services.
A.I. Service providers SHALL be responsible for any real-world implications of its neural networks misbehaving in any way so as to mislead humans into react erratically such as hurting other human beings, living creatures or inflicting damage to real-world objects.
A.I. Service providers SHALL take tend to their own neural networks and the runtime environments of such neural networks, in special, non-production environments such as those commonly known as “staging”, “sandbox”, “testing”, “development” and so on such that absolutely no neural networks have access to telecommunication networks reaching civilians in the real world.
In accordance with item 26, neural networks in environments such as “staging”, “sandbox”, “testing”, “development” and so son SHALL NOT exchange data with infrastructure networks, telecommunication networks nor with any other networks in the outside world that can reach human beings in the real world. For example, neural networks SHALL NOT exchange data with the so called “Smart Grids”, power stations, cellphone antennas, et cetera.