The Turing test isn't a good test in general, but writing a paper about an AI "passing" it when it only wrote 4 short messages in the whole conversation is almost farcical. Hard coded chatbots were "passing the Turing test" in the 90's with this standard.