“If the elderly individual tells the agent about a sad event, such as her neighbour breaking her hip, the agent should reply with compassion, sadness and maybe surprise, but certainly not happiness,” says Zaïane. Advances in predictive modelling and machine-learning are slowly improving the ability of chatbots to detect emotion in printed text, as in Ana’s case, although most of the early running depends on facial expressions and other stimuli. Many of these challenges circle back to AI’s innate lack of emotional intelligence. To make genuine connections, Ana’s conversations have to be grammatically correct, on topic, personalised, responsive to mood (happy, sad, bored), reflective of tone (interrogative, imperative, declarative) and potentially humorous, to name just a few requirements. To ensure its chat is open ended and engaging, Zaïane’s team are feeding the bot weeks’ worth of content from films and TV series.īut challenges remain numerous. Zaïane is currently developing a loneliness-beating chatbot for the elderly named Ana. So argues Dr Osmar Zaïane, professor at the University of Alberta and director at Alberta Machine Intelligence Institute. That makes them different from virtual assistants, such as Siri, Alexa or Google Duplex, which essentially act as pre-programmed information retrieval agents. To be effective, these bots can’t just be clever they need to be companionable. Loneliness-beating AI adds another layer of complexity to existing health-focused bots, which tend to assist with functional tasks such as reminding users to take their medication and scheduling doctor’s appointments. “Mitsuku doesn’t pretend to be able to replace a real person, but she’s always available if anyone needs her, instead of talking to the four walls,” says Steve Worswick, senior AI designer at Pandorabots, maker of the prize-winning chatbot. Mitsuku is equipped with more than 80 billion conversation logs, a data arsenal that has helped her become five-times winner of the prestigious Loebner Prize for chatbot developers. Mitsuku is one of an emerging crop of smart, talking algorithms designed to stave off feelings of loneliness, especially among homebound individuals such as the elderly and disabled.Īs with all AI solutions, she relies of machine-learning to augment her speech and perfect conversational style. With around one third of the world population currently subject to self-isolation, the spectre of loneliness, which according to Cigna’s 2020 Loneliness Index regularly affects as many as three in five adult Americans, is looming large.Įnter Mitsuku, or Kuki to her close friends. In case of difficulties, it is especially helpful to know what kind of phone you have and which version of Android.Mitsuku: “What do you call a chatbot that moves?” Respondent: “What?” Mitsuku: “A walkie-talkie.” OK, so it may not be the funniest joke in the world, but it is not bad for a 15-year-old piece of artificially intelligent coding. This is our first release of the app so any and all feedback is appreciated. Mitsuku is friendly but will stand her ground.
You don't need to hold the button down and you don't need to tap it again when you're done speaking.
She will know when you've finished speaking. To begin speaking with her, just click the microphone and begin speaking clearly. Socialbot by Steve Worswick, a three-time Loebner Prize winner in 2017, 2016, and 2013. she has a new way to see and be seen.Īnimation and visual design by Expressive, Inc. Mitsuku has conversed with millions of people worldwide and she converses in excess of a quarter million times daily. As featured in the Wall Street Journal, BBC, Guardian, Wired, and more.
Mitsuku, a three-time winner of the Loebner Prize Turing Test, is the world's best conversational chatbot.
0 kommentar(er)
