Have you seen the Spike Jonze’s film « HER » (2014)? As a reminder, the film takes place in the near future, Theodore is a sad man. He buys a sophisticated computer program which expresses itself as a voice, this program can adapt itself according to the user’s personality.
This program is named Samantha which is a feminine and intelligent voice. It appears to be intuitive and surprisingly funny. Samantha’s needs grow and evolve. Theodore discovers that this voice, created by the AI, is able of having feelings and emotions.
Can this situation happen in real life? We actually come closer to this situation. You already must have talked to a bot and you might have laughed during the very first conversations. While you surfed and tried to find out a piece of information on your favorite Internet operator site, suddenly, out of nowhere a window pops up and somebody is offering you some help. You first feel relieved not to be alone anymore, so you type the piece of information you are looking for in the window. But you are quickly disappointed as this “somebody” says it does not understand your request and therefore asks you to say it in a different manner.
Researches used to be like that, but not any longer…
A poet at the Silicon Valley
From now on you can count on writers and AI (Artificial Intelligence) conceptors who take the challenge to create the accurate approach between the bots and human beings. Cathy Pearl, Sence.ly’s VP says that users accept the bots’ mistakes as long as there is a sense of humor in the message. It is now common that in the Silicon Valley recruiters hire poets, writers, comedians and any person able to write any scenario elaborated by AI.
Their goal is to create a personality traits as real as possible during any computer assistance and to smooth the conversation with the internet user. Is it really feasible?
Writers will have to plan solid foundation for each interaction between AI and human beings. In order to anticipate potential answers, companies organize brainstorming sessions and make a review of all matters that can possibly be discussed about.
Facebook « chatbots »
Facebook announced on the 12th of April 2016 the creation of « chatbots » which should be likely to order a pizza or flowers via their messenger. Facebook offers a range of different bots. Any e-business company, transport company, service company or information will be able to create its own bot on Messenger. Uber, KLM or CNN are the first users of these bots. The bot of CNN can send the last news, can answer questions and adjust their topic proposals which interest you most while you are reading. A journalist from the American website Buzfeed has been the first to test this new Facebook system called M. He told how he got a ticket for the Star Wars premiere while his colleague had tried to find on internet a ticket during a whole evening. Facebook is not the only one: IBM is preparing with the Crédit Mutuel a “bank robot” that could meet a request such as receiving a check book or make a bank transfer either day or night.
Are robots really smart?
The irony of modern times is since human beings have created robots, they strive to prove that men will always be smarter than robots. But I am sure you remember this game of chess in May 1997 : Gary Kasparov beat the Deep Blue robot during the first game. But the Russian player lost the second round. He then lost his temper and left angrily the place.
More recently, we saw Tay, the Microsoft chatbot on Twitter has been “manipulated” by malicious users who have turned the robot into a nazi in less than 24 hours. When it comes to simple questions, everything is fine. It is more complicated when requests are sophisticated. Facebook promised to elaborate smart robots.
According to Forrester Data, chatbots will become inevitable in the working world. Circa 33% of the workforce will work with chatbots by 2019.