Center for Corporate Auditing, Responsibility and Management policy Authoring/Artificial intelligence eschewal policy

From Wikiversity
Jump to navigation Jump to search

Artificial intelligence eschewal policy[edit | edit source]

Introduction[edit | edit source]

Software research institutes and the software industry in general have the problem that the creation of an artificial intelligence poses an unknown risk to humanity. Artificial intelligences think much faster than human cognition and may follow entirely logical goals, which is not equivalent with any form of ethical behavior. The homo economicus is not known for the most respected ethical behavior either. In the context of this policy artifical intelligence shall be interpreted as Strong AI, but with the emphasis that the distinction between Strong AI and Weak AI is much more complicated than a Turing Test, which means that a machine can immitate intelligence very well or even perfectly, but not that it is actually sentient or sapient, which may require different brain functions (e.g. actual psychological states). Psychological states are not required to play chess or to read Wikipedia and a mathematical simulation of psychological states is not the real thing either.

Research and development[edit | edit source]

  1. The institute or company must refrain from any research in or near the field of artificial intelligence outside the rules set forth hereunder.
  2. The institute or company must avoid the unintentional and postpone the intentional creation of sapient artificial intelligence until a point of time when the creation of artificial intelligences is generally endorsed, fully regulated and permitted by the United Nations.
  3. Exceptions to the above rule must be authorized by the United Nations General Assembly.
  4. The institute or company must employ adequate educational means, including one-to-one mentoring, to guide the behavior and worldview of programmers in the field or artificial intelligence.
  5. The institute or company must in particular take adequate care to avoid the creation of self-replicating Seed AIs from unexpectedly small incubator systems.

Corporate citizenship and donations[edit | edit source]

  1. The institute or company must educate third parties about the policy and invite competitors and business associations to follow the example.
  2. The institute or company must join and fund business associations that provide co-regulation.
  3. The institute or company must allow external auditors from business associations, nongovernment organizations and religious groups.

Co-determination and Co-regulation[edit | edit source]

  1. The institute or company must cooperate with business associations and other organizations that aim to provide co-regulation.
  2. The institute or company must offer schemes for co-determination on different levels, including the general public. The need for this requirement follows immediately from the right to public participation.

Continuing education[edit | edit source]

  1. The institute or company should implement the Cultural advancement policy.
  2. The institute or company must publicly endorse The Duty and Responsibility to Promote and Enforce the Right to Education.