may · contain · language

This is Elbot. Elbot fooled 3 out of 12 judges in the 2008 Loebner Prize contest. There are articles about it all over. This one, for example. At least two Swedish sources seem to make the claim, or at least imply, that this is a better result than any robot has gotten in a Turing test, at least while competing for the Loebner Prize. It seems pretty clear looking at the different articles that there has been no better result, but I'm having some trouble finding sources for Elbot being the first to manage 3 in 12. So I won't say that, but I do think that's how it is.

If that wasn't clear, you can chat with Elbot here. It's sort of fun.