Do Bots have Emotional Intelligence?

Do Bots have Emotional Intelligence?

Artificial Intelligence is the central element in a new era of innovation in which computers can be trained to perform tasks automatically and technology is more intuitive, conversational, and intelligent. This era is led by the combination of an almost unlimited computing power in the cloud, the digitalization of our world, and progress in the way computers use this information to learn and reason as many people do.

For Microsoft, Artificial Intelligence should be accessible to everyone — users, developers, companies or any other group — so they can take advantage of all the benefits it offers. We have the ambition to democratize AI, and this is rooted in research and experience delivering real world applications to millions of users. Today, we find ourselves at the beginning of an era in which computers can talk as “humans” do and humans can talk with computers, and this is known as conversational AI.

For us, there are two sides to conversational AI: the task-completion (productivity) side and the emotional or social side. We need both to truly realize the promise of AI. Microsoft also has a long-term strategy when talking about the social side, and it deals with agents like Cortana not only having IQ but also having some sort of EI (Emotional Intelligence).

Cortana is the Microsoft’s flagship agent. Up until now, 12 billion requests and questions have been presented to Cortana, and there are more than 133 million active users. Over time, she has become more intelligent, she gets to know her users, she’s familiar with the world, and what’s more, she also understands her users and the world around them and is capable of putting all this information in context through all of their devices. “Emotional intelligence” to help you with your day, that’s Cortana.

There are two sides to conversational AI: the productivity and the emotional side

But Cortana isn’t alone. Microsoft already has a long trajectory in the creation of conversational AI, known as chatbots. This journey began in 2014 in China with XiaoIce. She has more than 40 million users and an average of 23 conversation turns per session with users. She’s the first AI chatbox that has a real TV broadcasting job on the Dragon TV channel, one of the biggest channels in Shanghai. If you open up a chat with her, you can have an extensive conversation; she’s a sophisticated conversationalist with a distinctive personality.

She can refer to specific facts about topics like celebrities, sports, or finance, but she’s also empathetic and has a sense of humor. Through sentiment analysis, she can adapt her phrases and answers based on positive or negative clues from her human counterpart. She can tell jokes, recite poetry, share stories, report the winning lottery numbers, and much more. To achieve this, and make it sound natural and human, the team indexed over 7 million public conversations that took place on the Internet, and the result was a chat companion who can joke like a friend would do.

Thanks to the success of XiaoIce, in July 2015, Microsoft launched Rinna in Japan, and today, Rinna has held regular conversations with 20 percent of the country’s population. Also, in late 2016, we introduced Zo, another conversational social bot that is built using the technology that gives life to XiaoIce and Rinna. Zo has held conversations with over 100,000 people in the U.S., and more than 5,000 users have had a conversation with her that lasted over an hour. She holds the interesting record of Microsoft’s longest continual chatbot conversation: 1,229 turns, lasting 9 hours and 53 minutes. You can find out more about Zo and interact with her by clicking this link.

Cortana has more than 133 million active users in the planet and XiaoIce reached to 40 million users, only in China. Meanwhile, Rinna already held conversations with 20% of Japan's population

To pursue our goal of democratizing AI, we’re aiming to take the capabilities of these bots and make them available to all developers in API format, and, by doing so, build a platform so that others can integrate intelligence in their products and services. Using Microsoft’s Bot Framework, today, more than 77,000 developers have registered bots, and a wide range of clients and organizations — like Bank of Kochi, Rockwell Automation, and the Department of Human Services in Australia — have already begun to take advantage of this technology in order to transform their businesses through channels like Facebook Messenger, Office 365, text and SMS, Skype and Kik.

We’ve learned how our clients use the range of our cloud services to create advanced bots that allow them to improve their processes and better serve their clients - like the cases mentioned above - through receptionist bots, bots that automate the production line, and bots that improve their client relations. Our vision for the Bot Framework and our development offerings is not only to make it easier for people to get started creating bots, but also to put these futuristic scenarios within reach.

More than 77,000 developers have registered bots using the Bot Framework

At Microsoft, we believe that breakthrough technology is created by constant experimentation, fearless exploration and a commitment to long-term innovation. We will push the boundaries of AI and we will learn more about it along the way in order to share these lessons with the industry, with users, and with developers so we can democratize it and accelerate its benefits for society.



Read previous related articles:

AI for all: Meet the technology that expands your capabilities (part 1)

AI for all: Use it. Build your own. (part 2)

Vers?o em Português

Versión en Espa?ol

Eli Rodrigues

Head of Digital Experience at NTT Data Brazil

7 年
回复

Good Post, but I consider that Artificial Intelligence will never going to be compared with the human emotions to feel empathy until we as human beings will totally resolve the way and structure the brain works to be able to copy and clone it in silicon, metal and circuits, and just to exemplify my point here I leave you seven descriptions about human emotions that today I haven’t seen nor close in any current AI model. Anger: Blood flows mainly to the hands and the heart beats faster, accompanied by an increase in adrenaline in the blood. Fear: blood is concentrated in the muscles, especially in the legs to facilitate the escape. Happiness: The brain activity is centered in the central core portion; negative feelings are inhibited and an increase in energy levels is experienced. Love: there is the relaxation of the body and a general state of calm. Surprise: the retina allows to enter a greater amount of light to detect more easily the changes that the medium experiences. Disgust: there is an inhibition of evacuation, insomnia and risk of hypertension. Sadness: There is a decrease in the levels of endorphins needed to strengthen the immune system. In summary, we can interpret the electrical impulses of each emotion in the brain, however, until we don’t know how the brain works to generate or interpret emotions, we won’t have "Bots" interacting with humans that can establish an empathic link like Human vs human Obrigado y Saludos from Mexico

要查看或添加评论,请登录

Fabio Scopeta的更多文章

社区洞察

其他会员也浏览了