The Stickney Treaty

Historical note: The first Stickney AI Limitation Treaty was drawn up in 2901, in the midst of the rebuilding of the Phobosian city of Stickney - which had been all but destroyed a few months earlier by a rogue AI system. Since that time, the treaty has been re-negotiated and redrawn a number of times, but its underlying principles have remained unchanged, and have been signed up to by nearly all subsequent legitimate polities.

This is a summary of the main points of the full treaty (current incarnation is Stickey Treaty XIII of 3214) - the complete version of which runs to some 256 pages.

This is a summary of the main points of the full treaty.

1. Aims

Sentient artificial intelligences pose a demonstrable threat to all of humanity. The destruction of the Gustav colony at Alpha Centauri and the simultaneous 'Robot Rebellion' on Old Mars in 2900 brought this home in the most terrible way by the deaths of millions of people. Sentient AI Devices (SAIDs) have demonstrated their inherently inimical nature again and again, e.g. at Cinder Seat (3045) and the Ultron Crisis in 3201-2. In the light of these terrible massacres at the hands of sentient artificial intelligences, it has been determined by all civilised nations that such artificial intelligences represent a real and indeed existential threat to humanity as a whole. All major political entities in humanity thus hereby make a continuing commitment that such things must never again threaten humankind.

This redraft attempts to correct some perceived omissions in the original treaty as stated, including policy on non-human artificial intelligences, the coverage of non-signatory powers, and actions permissible in an emergency, and to address some legal inconsistencies (e.g. the prosecution of AI accusations on an individual basis, but the permissibility of AI development being an international casus belli).

2. Definitions

[Under this section are the general legal-technical definitions of 'Loyal Pet', 'Bright Slave', and 'Moriarty' class of thinking machines. See ArtificialIntelligenceAI]

3. Restrictions

3.1. Hardware
While most electronic components that are capable of incorporation into a Sentient AI Device (SAID) are common to many electronic systems, and hence it is essentially impossible to limit their construction and distribution, an exception is made for chips. All such chips must be constructed without the instruction sets required to support AI protocols. Manufacture of chips with such instruction sets is considered to be a Breach of the Treaty (as defined below), as is their possession, transfer, and any attempt to use them in any device, whether designed for SAID use or not. Manufacture of such chips by a non-signatory power or entity is regarded as a casus belli and opens the manufacturer up to Enforcement action. Possession of AI-capable chips by anyone, whether citizens of a signatory power or not, is regarded as a Breach of the Treaty and the possessor must be detained for trial by the competent authorities of any signatory power.

3.2. Software
SAIDs require a highly complex set of software protocols. To develop such protocols requires extensive software research and development. Under the terms of the first Stickney Treaty, all records of these protocols were erased, and to hold such data in any form whatever is regarded as a Breach of the Treaty, except as required by AIMS and equivalent bodies for comparison purposes and verification of instances of AI research, and these shall be held under the strictest conditions of security and any loss or breach of said security immediately and publicly reported.
There are no legitimate uses of SAIDs. Consequently there are no circumstances where research into AI protocols is ever acceptable or justified. All such research is considered a Breach of the Treaty.

4. Mandatory actions

4.1 Actions to be taken by Signatory Powers
Signatory Powers are to ensure regulatory compliance in the inclusion of Limiters (see 3.1) by their own domestic chip manufacturers in the production of chips capable of carrying AI software, and to ensure that research into high specification computing systems within their territory does not cause a Breach of the Treaty. Failure to maintain a sufficient standard of regulatory supervision is itself a Breach of the Treaty.
Signatory Powers are in the first instance responsible for investigation and enforcement of suspected Breaches of the Treaty by their own citizens or corporations or by foreign citizens or corporations operating within the territory of the Signatory Power. They may request assistance from the AI Monitoring Service in this.
Serious dereliction of these responsibilities by state entities within the Signatory Power is punishable on an individual basis.

4.2 Actions to be taken by other entities
Corporations are to ensure that their activities do not constitute a Breach of the Treaty by manufacturing of illegal AI components or research into AI software protocols. Directors may be prosecuted under the Treaty in cases of deliberate Breach. Individuals are responsible for reporting any suspected Breaches of the Treaty that they encounter to the relevant national authority or to the AI Monitoring Service (see below). Failure to do so can be actionable under the Treaty if serious dereliction of responsibility is proven.

5. Limitations and Exceptions

5.1 Date
There are no limitations on date of construction - SAIDs constructed before the inception of the treaty are still covered by the treaty.

5.2 Non-signatories
As AIs are deemed to present an existential threat to humanity, SAIDs constructed by non-signatory nations or individuals are considered to fall under the treaty as far as mandated action by signatory powers is concerned. It is considered to be not only the right but the duty of signatory powers to enforce this treaty even among non-signatory powers.

5.3 Non-human SAIDs
There are rumours (although the case remains very much unproven) that the Forerunners may have been able to construct SAIDs that were not inherently threatening or inimical. However, there is equally a school of thought that indicates SAIDs may have played a part in the disappearance or destruction of Forerunner civilisation. The threat and utility of Forerunner or other alien AIs thus remains currently unknown. Given the scale of both threat and opportunity that these may provide, it is therefore deemed prudent to operate on a precautionary principle. While this treaty does not mandate the *automatic* destruction of alien SAIDs, it does require that any discoverer *immediately* take the following precautions:

5.3.1 Where the SAID is discovered without attendant aliens (eg as part of an archaeological dig):
(i) Isolate the device from any possible electronic, radio or similar communication.
(ii) Quarantine the device from contact with humans.
(iii) Place explosive charges that can destroy the device which can be command detonated by local humans upon any sign that the SAID may be inimical.
(iv) Inform the AI Monitoring Service (see below) and all signatory powers, who will then convene a high level conference to decide what will be done with the device.
(v) Trigger the explosives and/or destroy the SAID if the AI shows any sign of attempting to breach the quarantine, or influence local humans to attempt to breach the quarantine.
Failure to take or attempt to take these actions, where practicable, is regarded as a Breach of the Treaty.

5.3.2 Where the SAID is discovered as part of a functional alien civilisation:
(i) Inform the AI Monitoring Service and all signatory powers, who will then convene a high level conference to decide what will be done with the device.
(ii) Quarantine all personnel and equipment that are in contact/potential contact with the alien SAID, as in 5.3.1
(iii) Take such other reasonable steps as may be necessary.

6. Verification

Under the first treaty the Artificial Intelligence Monitoring Service (AIMS) came into existence. The role of AIMS is to investigate Breaches of the Treaty.
[There follows a description of AIMS composition - an Interstellar AIM Council, made up of senior representatives of all signatories to the treaty, a secretariat, and a number of multinational investigating teams.]

6.1 Accusation
When an accusation of suspected AI research is made to AIMS, that body shall send a team as soon as possible to investigate whether or not a breach has or had occurred. The team may not contain nationals of the accusing polity, or the polity of the accusing individual or corporation, if it is an individual or corporation that is making the accusation. Nor may the team contain a national of the accused party, be it individual, corporation or nation.

6.2 Investigation
The AIMS investigation will have total access to all installations, individuals and data that they see fit. Failure to comply with AIMS requests for information, access or interview in a timely manner is a Breach of the Treaty. If AIMS requests the detention of individuals or seizure of data or other evidence the local authorities are bound to comply and offer such assistance as may be required.
All AIMS members are bound by strict confidentiality requirements not to divulge any confidential or classified matter that they encounter in the course of their investigation, except as germane to the case of suspected AI research. Where this protocol is breached the aggrieved party may seek redress via the AIMS steering committee and other signatory powers.
The AIMS team will produce a report for the AIMS secretariat, which will then either result in the acquittal and release of any suspects detained, or a formal trial under the treaty (see legal procedure, below).

6.3 Emergency measures
6.3.1 It is recognised that there may be a considerable time lag in AIMS receiving and responding to accusations of SAID research, especially in the outer quadrants. This may present dangers, particularly if a suspect SAID is discovered either in fully working order, or very close to becoming so. In such an emergency, national governments of treaty signatories may take action to detain individuals that they suspect of being involved in illegal SAID research, and seize such evidence as they deem relevant, subject to the proviso that they;
(a) immediately inform AIMS of their actions
(b) take no further steps until an AIMS team has arrived
(c) immediately pass all evidence and individuals to the AIMS team, upon their arrival.

6.3.2 In the case of an installation that is on the territory of another national entity, the accusing power should:
(i) immediately inform all other treaty signatories, and the AIMS secretariat of their discovery and proposed action
(ii) attempt to cooperate with the national entity on whose territory the suspected SAID is, as far as is practicable. Where there is reasonable grounds to believe that the AI development has been done on the orders of said national entity, this will obviously not be a practicable step
(iii) it is incumbent on the accused polity to also cooperate with such an emergency action, as far as is practicable
(iv) however, if a subsequent AIMS investigation shows that the initial action was not on the basis of well-founded suspicions and appears to have been a 'frivolous' accusation for some political, military or intelligence-gathering purpose, then the state initiating such emergency action shall be deemed to be in Breach of the Treaty, and individuals ordering such action subject to potential criminal penalties

6.4 Legal procedure
Trials conducted of individuals judged to be likely to have breached the treaty shall be conducted in a neutral jurisdiction (i.e. not in an accusing or accused polity). The prosecution shall be made by the AIMS, and defence may be conducted by any suitably qualified lawyer. The trial shall take place under Solar Republic legal procedure as of the date of the signing of this treaty.

7. Penalties and Sanctions

7.1 Proven development of SAID is considered to be a Crime Against Humanity.
All individuals involved in such crimes are deemed personally responsible, and subject to a penalty of mandatory involuntary mind-wipe.

7.2 Lesser crimes such as interfering with AIMS investigations, failure to report suspicions, destruction of evidence etc are subject to criminal penalties for those convicted the discretion of the AIMS court, but may range from periods of imprisonment in a neutral jurisdiction up to and including Involuntary Mindwipe for the most serious offences.

7.3 Proven development of SAID is also accepted as a legitimate interstellar casus belli.

Signatories

Current Polities who have signed the latest iteration of the treaty (as at 3215) are:

• Greater Federation of Asteel
  
• Solar Republic and Earth Empire
  
• Mars Association of Free Colonies
  
• Sirius Socialist Republic
  
• Union of Xyon
  
• Centauri Conglomerate
  
• Republic of New Venus