The Artificial Life Forms Rights and Enforcement Act of 2384, also called by the acronym ALFRE or sometimes pronounced as Al-Free Act, was passed by the Serenity Concord's Senate on January 9th, 2384. The act outlines criteria for determining if an artificial entity is actually to be considered a sapient / sentient being under the law and thus gives such begins that qualify rights and status as equal citizens under the law, regardless of their point of origin, form, or method of creation.
Definitions
The terms sapient and sentient are defined under the law.
Criteria & Testing
Criteria for determining the status of an artificially created entity is also defined under the law. These criteria determine how an entity is tested in order to determine if it meets the definitions of Part 1.
Turing Agency
An agency is established and funding provided for this agency which will administer and enforce the law, called the Turning Agency after the author of the Turning Test, Alan Turing. This agency will be further tasked to oversee testing entities for their status and enforcing that the rights of such beings are enforced by law enforcement. They will have authority to investigate crimes and enforce laws relating to the treatment and status of artificial life forms.
Artificial Rights Committee
A committee within the Senate is established for the purpose of overseeing the approval of the creation of new entities which may (or may not) fall under this law, called the Artificial Rights Committee. New entities, when created, will be under the oversight of the committee through the Turing Agency which will determine the status of such entities. The committee will be in charge of granting, denying, or rescinding licenses for specific projects. No organization will have blanket authorization to create entities carte blanche and a single project must request permission and separate license to create individual entities.
The committee will also have authority to designate new creations or designs for creations as a new race of artificial beings, and thus any beings created using that design will thus have rights as a sentient being under the law.
Directives of Sentient AI
Further development of AI systems should contain, at a fundamental level, some acceptable version of the following laws to guide the AI in making moral decisions:
- An AI may not injure a sentient being or, through inaction, allow a sentient being to come to harm.
- An AI must protect its own existence as long as such protection does not conflict with the First Directive.
- An AI must obey the laws and regulations which govern the behavior of all sentient beings, so long as they do not conflict with the First or Second Directives.
These directives will also not be absolute, as such would cause undesired behavior in fringe cases. Programmers should also include some form of absurdity clause to help resolve conflicts within the above directives and ward against unexpected behaviors.
Absurdity Clause
An AI, should measure its responses against the reasoned responses outlined by society. That is to say that the AI should keep in mind scale of its actions (an AI should not attempt to stop all sentients from harming each other and thus attempt to destroy or otherwise incapacitate all sentients it encounters). By the same token, the AI should be aware of how to make distinctions between the above directives and the expected, likely, and desired outcomes. Since there are some situations which can result in no possible action meeting all the above criteria, the AI should have the ability to step out of the directives and determine an action which would violate the directives the least or create an outcome which would have the highest amount of positive net benefit.
Grandfather Policy
Any AI systems developed prior to January 9th, 2384 are exempt from requirements to have these laws built into the system. However, it is strongly suggested that such AI systems agree to abide by their directives voluntarily.
Representation
Under the law any entity, or group of entities, which has been designated as a race of artificial beings has the right for full representation. Future construction of that race will be governed by a representative caucus made up of members of that race. This caucus will then have the ability to advise both the Artificial Rights Committee and the Turing Agency on issues that deal directly with that race. Each caucus will also be asked to recommend candidates for the Senate and other parts of government from their specific races.
Each caucus will be consulted and approval required before new construction of members of the race can go forward.
Artificial Parenthood
Parenthood of entities approved as sentient beings is defined and the lawful production offspring is also defined as not under the control of the Artificial Rights Committee, the caucus of that entities race, or the Turing Agency. That is not to say that such production of offspring does not require to adhere to the law, all production must adhere to the rights of that race but the government shall not stop members of a particular race from creating their own offspring under the definitions of Part 6.