Strong AI

“Strong AI” is also known as “General AI” and refers to Artifical Intelligence that meets or exceeds human intelligence as opposed to AI that is designed for a smaller or specific purpose. It is generally assumed that this will include:

  • Communication ability
  • Knowledge representation
  • Reasoning ability
  • Learning ability
  • Ability to integrate all of these

The benchmark for determining strong AI will need to go beyond the Turing test which can be met with a sophisticated mimicry of human behavior. It is possible that the standard for strong AI will become for the AI to be able to be considered more human-like than most human-beings. Communication, reasoning and learning ability can be defined and programmed and tested. The key issue is determining to what standard these areas will be tested. Knowledge representation and overall integration can also be defined, programmed and tested but setting standards and defining semantics for these areas are more difficult.

While historically, strong AI has been associated with functional automatons or robots that are capable of initiating human-like behavior, there are other forms of strong AI that may precede the “intelligent robot”. As AI based “advisor” programs become more competent in specialized fields, the need for them to consult general fields of background data will grow.

Turing Test
Beyond Turing
Self Awareness

Comments are closed.