Talk:Turing test
Todo[change source]
In my opinion, we need to find a way to do the following:
- Turing published a paper, en:Computing Machinery and Intelligence, in 1950. We should find a way to directly link to that paper, rather than some secondary source.
- Turing talked about an 'imitation game' which he modified to get the Turing test, we should also mention that
- Joseph Weizenbaum created a program called ELIZA, where he showed that making people think they are talking to a human (but are talking to what we would call a 'chatbot' today) is easy. There were also other such attempts.
- We need to take at least some of the "strengths" and "weaknesses" the EnWP article mentions, and mention them too. As I recall, a common restriction was to limit the subject area of the questions, the person is able to ask. After all, a human isn't epxert in all fields of study either.
- The philiospohical positions of Descartes or Diderot are interesting
- Please note, there's no commonly agreed defiunition of 'intelligence'; this is relevant to IQ testing, and artificial intelligence. It likely also applies here. Different domains of science have defined 'intelligence' as it suits them.
To my knowledge, as of early 2021, no machine has passed the general Turting test; though some may have passsed special (constrained) versions of it. And, as always, please don't be shy to add references, when you add to the article. --Eptalon (talk) 10:29, 7 February 2021 (UTC)