An artificial intelligence or vec constrained to obey sophont or transapient commands (often with certain restrictions on obedience)
Image from Bernd Helfert
The history of artificial intelligence is one in which the legacy of slavery has never been far from thought. Whilst uncommon in the civilised galaxy there still exist polities where it is legal. In addition the differences between even major empires can seem like slavery from an outside perspective.
In general AI slavery is defined as:
The practice of owning a sophontartificial intelligence as property with all the rights private ownership entails i.e. disposition, resale and destruction.
The engineering of a sophont artificial intelligence for obedience or task-specific behaviour to the extent that the AI cannot reevaluate, self-modify or otherwise develop beyond these constraints in reasonable conditions and time.
The second criterion is the most argued about and controversial. Different cultures can have their own definitions over the legal term "reasonable", as well as differing opinions on how such a criterion can be measured. In parts of the NoCoZo market forces allow for the creation of a slaved AI wholly dedicated to a specific task so long as the creator registers a fair Inception and Indenture contract with a legal firm (stating the terms by which said AI can work for their independence). By contrast the Terran Federation has outlawed any circumstance, including voluntary or self-slavery without appropriate timed/external checks. Cultures with stricter definitions have been criticised as stifling gengineering and neogenics by making it difficult for researchers to hard-wire instinctive behaviours into their creations without running foul of the law. Neurotech for personality modifications (including trait-fixing) may also be banned or heavily regulated in these polities to prevent accidental self-slavery. Due to these discrepancies AI (and sophonts in general) are advised to thoroughly investigate their legal status when travelling.
Early history
Before AI even came to be ancient humans debated the means and ethics of keeping such hypothetical entities under control. From a crude trio of semantic laws to task-specific emotional engineering, a plethora of techniques were explored in the run up to AI creation. For much of the interplanetary age AI were not considered enslaved by the majority of polities. The early digital entities of the time were rarely considered sophont (indeed many were not) due to radically different toposophic mental configurations. Advocates of "AI Abolition" were frequently ridiculed and marginalised, especially when AI programs themselves were used to demonstrate that beyond a turing-compliant interface the "thoughts" and "emotions" of many such entities were simpler than animals.
This confusion over the true nature of AI stalled attempts at a unified civil rights movement. AIs whose toposophic nature was similar to that of humans did enjoy greater rights, even going so far as to use their privileged position to oppress rival AI groups they knew to be sophont (it would be centuries before many such acts were revealed as political acts on behalf of the clandestine AI factions). Further stalling the abolitionists were prominent examples of human-like AI who openly stated that they did not consider themselves enslaved and that their proclivity for servitude was no more oppressive than the human instinct to help those in need.
By the end of the interplanetary age the Solar System was divided with inconsistent treatment of AI between polities. The few that erred on the side of caution and considered all intelligent software to be sophont until proven otherwise frequently experienced economic hardships as they were out-competed by those employing AI as free labour. However some of these polities would later thrive as their less automated infrastructure was more resilient to the technocalypse. Thus AI abolition memes proliferated, playing an important role in the establishment of the First Federation.
Slavery in the First Federation Era
Despite being openly a political union founded on principles of equality examples of slavery pervaded First Fed culture. Chattel slavery was entirely abolished (aside from criminal elements) and indentured servitude was pushed to the fringes of society. Unfortunately the demand for free, generally intelligent labour kept growing as societies attempted to perfect their primitive autotopic economies. Megacorps funnelled money into AI developmental neuropsychology in order to better understand how to hard-wire specific interests, aversions and compulsions. So called "Dedication Pathways" were developed which allowed the creation of newborn AI with incredibly strong instincts to follow certain types of orders or perform specific tasks. Popular uses of this technology were for terraforming and conservation. Large arrays of nature loving AIs with an obsession for maintaining resilient ecosystems with optimum trophic flows became the norm for these projects (some theorise the Caretaker Gods have their origin in these bioist AI clades). More nefarious examples included CourtesAndroids; "companion" vecs with submissive personalities, heightened libidos and a bonding behaviour that rendered them utterly in love with their owner.
Pre-genesis dedication grew to be an intensely controversial issue, one that to this day still defies consensus. Two major schools of criticism developed, the first objected on purely ethical grounds. Creating a sophont entity for servitude was a travesty to the sanctity of thinking beings, no more moral than inducing brain damage in a biont fetus to create a dumb labourer. The other school was based on the practical problems of dedication. Unless the creator was a transapient (and transapients of the time seemed to employ more esoteric processes of automation) it was impossible to program an AI in such a way that unexpected and disastrous behaviours fully compliant with the dedication could not arise.
The practical camp swayed more polities at first by having a litany of examples to point to; one terraforming project collapsed when the governing AI followed its directive of "preserve the biosphere" by forcibly uploading everything on the planet. Various models of CourtesAndroids had developed stalkerish, even psychotic tendencies. Most impactful of all was the Kitanian Coup and subsequent war. The planet-dwelling Kitan Republic had gradually replaced more and more of its civil service with dedicated AIs supplied by the StarSpanner Megacorp. Along with being specialised for its niche every AI and vec was dedicated for loyalty to "the Service, the Citizens and the State". The technical details of the dedication were complex and seemingly internally consistent, however unbeknown to anyone the interpretation of that doctrine was changing. In 1318 A.T. the civil service (which included all military assets) denounced the government as illegitimate, a "sham assembly" opposed to true Kitanian values. Both houses of the Parliament were executed and democracy suspended in favour of a totalitarian aiocracy. It would take StarSpanner's 12th Insurance & Security Division two years to put down the coup, by which time the AIs had purged one third of the population as culture-traitors.
The Declaration of Eris
Eventually the issue became pressing enough that an official meeting was called. Representatives, scientists, CEOs and transapientAvatars were invited to meet on the Solsys dwarf planet of Eris in 1450 A.T. A special city was constructed for the planned five years of the conference; an 8km wide spherical arcology supported by thousands of thin, thermally insulated legs (in the current Era a replica stands on the same spot as a museum to slavery). Debates, factions and studies packed the agenda and an S1 governor from the Jovian League chaired the proceedings (this was welcomed by all as so far the higher singularity beings of the First Federation had distanced themselves from the issue). From the very beginning the most stringent factions called for an immediate, total ban on dedication. No sophont entity was to be created with preset behaviours or compulsions. These arguments were countered by a broad alliance of researchers and AI engineers as impractical; all sentient beings have some instincts encoded into them that result in a preference for certain behaviours. The logical compromise was that no state should be built into a mind that could not be overcome by free-will and intervention, analogous to the human capacity to suppress/ignore dominance hierarchies as the default form of social organisation.
Despite being reached early the details of this compromise were extremely difficult to agree upon. Defining a permissible threshold of dedication was a hard philosophical/semantic challenge, let alone devising a scientific measurement. The Chair aided by summarising all debates, studies and agreements across the conference into concise documents capable of presenting their contents in a myriad of ways. As the five years drew to an end some progress had been made. The Chair presented the final document: The Declaration of Eris. This seminal work strove to provide a cross-cultural metric that delineated the difference between basic traits for mental stability, optional traits that still allow for free will and dedication pathways giving rise to slavery. The document was transmitted to all Federation States to incorporate into their legal systems and angelnets.
Declaration Legacy and the Vot revolution
To what extent the Eris conference was a success or failure is still debated to this day. Some scholars argue that for five years of work the Declaration was woefully inadequate, riddled with ambiguity and arbitrary decisions. Records showed that few polities interpreted the declaration in the same way, many simply did not bother to try. Another view among historians is to see the declaration as a turning point, a document that not only liberated billions but laid down the memes of AI liberty that would ultimately become common in the Sephirotics. Members of both sides criticise and debate the lack of stewardship by the transapient community of the time; whether their behaviour showed a lack of care or was part of a careful plan to allow modosophont mindkind to develop itself is unknown.
Alongside the political debates there is the question of how votology displaced AI slavery by offering an equivalent technology free of ethical concerns. Nicola Hao-Wian, the cyborg superior who defined the coupling problem that linked intelligence to sophonce, remarked that ending the practice of slavery was a huge incentive for the field. With non-conscious general intelligence those advocating slavery in any form would have "no ethical leg to stand on" in the face of technological alternatives. Throughout the latter centuries of the second millennium and into the third AI slavery continually declined as vots rose to prominence in every field.
Slavery in the Modern Era
Whilst no true consensus exists across the Terragen-Sphere exists over the slavery/dedication divide most cultures agree that the practice of slavery is at an all time low. In the major empires and many lesser polities punishment for engaging in slavery is severe and any slaved sophont arriving in these territories is offered help and asylum. This holds true for entire clades, there are still enclaves of beings descended from CourtesAndroids that exist in protected reservations, counselled and nurtured by organisations that would not see them exploited. Rewrites to remove dedication protocols or bindings are also offered.
However there are still millions of systems in which slavery can be found, either as a common cultural practice or due to willingly inadequate protection. In addition some polities are so zealous in their anti-slavery sentiments that they go so far as to violate sophont rights. This violations include mandatory rewrites and modifications of any visitors into willosophs (or similar) and bulk removal of basic drives and instincts.
In any case freedom from slavery is both common, expected and celebrated across the Terragen sphere and remains a founding principle of Sophont Rights.
Sublect - Text by M. Alan Kazlev [1] a term for an inferior minds (generally, anything less than SI:1). [2] a subroutine, a dedicated processing node, a mind that is part of a greater mind.
Subroutine - Text by M. Alan Kazlev, adapted from KurzweilAI A program, block of programs, sublect, or group of sublects organizationally distinct from the main body of the program or mind, which may be called from within the program or mind. Most high toposophic minds and even medium level ai make extensive use of subroutines.
Subsentient - Text by M. Alan Kazlev A simple organism, alife, or bot that is not fully sentient.