Chatbots are evolving beyond simple question-answer systems. They are learning to engage in small talk, a significant development in artificial intelligence (AI). Technologists are programming chatbots to understand and respond to idle chit-chat, a task more complex than it seems. This is because small talk requires understanding context, mood, and subtlety, which is a challenge for AI.

Engineers at Microsoft’s Bing are using a method called ‘sequence-to-sequence learning’ to train chatbots. This technique is based on machine learning, where the chatbot learns from previous interactions. The more the chatbot interacts, the better it gets at understanding and responding to small talk.

The potential applications of chatbots are vast. They can be used in customer service, where they can handle routine queries, freeing up human agents for more complex tasks. They can also be used in healthcare, where they can provide mental health support.

Despite the progress, there are still challenges. Chatbots sometimes generate inappropriate responses due to their inability to understand context fully. There is also the risk of over-humanising chatbots, which can lead to unrealistic expectations. Nonetheless, the evolution of chatbots signifies an important step forward in AI.

Go to source article: http://phys.org/news/2016-05-future-chatbots-small-talk.html