In the world of AI, the quest to meet the high hurdle of passing the Turing test is historic. To date, no-one has claimed the $100,000 grand prize for meeting all 3 requirements or even 2 of the $25,000 prizes for meeting any one requirement.
This Year’s Winner
Each year since 1991, a smaller, $2,000 - $4,000 prize has been awarded to the AI program that comes closest to beating any one criteria of the Turing test. This year will be the first time that one person has won a Loebner prize for a 5th time and the first time it has been won 4 consecutive times by the same person with the same program.
This year, that prize belongs to Steve Worswick and his increasingly humanlike Mitsuku chatbot.
Meet Mitsuku
Mitsuku is an incredibly human-like chatbot that emulates the personality of an 18-year old woman in Northern England. Anyone with a web connection can freely chat with Mitsuku and ask her anything at all and she will respond. She (it?) does a good job emulating human responses, recognizing modern colloquialisms, computer-abbreviations like LOL, and current events. Ask about her favorite football team, and she’ll ‘happily’ tell you about Leeds United. In an ironic twist, ‘Terminator’ is her favorite movie.
While Mitsuku has been improving year over year, there are still recognizable differences between the chatbot and a human on the other end of the text. Mitsuku has a hard time recognizing common misspellings or transposed letters, and conversations have a finite endpoint even if the person interacting with ‘her’ tries to continue the conversation.
The Quest Continues
Mitsuku did its best to fool the 2019 Loebner Prize judges, even despite the program’s willingness to respond to a specific query in which she will admit to being AI. And, interestingly, this year there was a twist; the test was integrated into a public event on AI, allowing over 200 children to interact with the AI chatbots. A total of 17 bots from 8 countries were featured at the event.
Still, none of the AI entered into the contest was successful at fooling humans in all three categories. So ends another year without a grand prize or runner-up Loebner prize winner. However, all is not lost, and in fact, far from it. Each year that goes by, the contestants bring the world closer to an AI system becoming so advanced that it passes the Turing Test. We’ll just have to wait and see…what will happen in 2020!
