AI-complete

English

Etymology

By analogy with terms like NP-complete.

Adjective

AI-complete (not comparable)

1. (artificial intelligence, informal) Of a problem: such that a solution presupposes or entails solving the problem of constructing a general artificial intelligence ("strong AI").
   • 2006, M.Gori, M.Ernandes, G.Angelini, “Cracking Crosswords: The Computer Challenge”, in Oliviero Stock, Marco Schaerf, editors, Reasoning, Action and Interaction in AI Theories and Systems: Essays Dedicated to Luigia Carlucci Aiello, Springer Science & Business Media, →ISBN, page 266:

     Problems like solving crosswords from clues are reputed as AI-complete [7]. This enormous complexity is due to its semantics and the large amount of encyclopaedic knowledge required.

   • 2007, Ben Goertzel, Cassio Pennachin, Artificial General Intelligence, Springer Science & Business Media, →ISBN, page 452:

     Just as concept kernels are not AI-complete, sequiturs and resonances are not AI-complete. Sequiturs and resonances also may not need to be human- equivalent to minimally support deliberation; it is acceptable for an early AI to miss out on many humanly obvious thoughts, so long as those thoughts which are successfully generated sum to fully general deliberation.

   • 2014, Nick Bostrom, Superintelligence: Paths, Dangers, Strategies, Oxford University Press, →ISBN, page 71:

     ... Now imagine a remote descendant of such a system that has acquired the ability to read with as much understanding as a human ten-year-old but with a reading speed similar to that of TextRunner. (This is probably an AI-complete problem.)

Translations

such that a solution presupposes constructing a general AI

Further reading

• Eric S[teven] Raymond, editor (2003 December 29), “AI-complete”, in The Jargon File, version 4.4.7.