Experts attending the recent DLD ’17 conference in Munich predicted bots – software apps which conduct automated tasks – will pass the Turing test within a decade, and so become a key new tool in customer relationship management (CRM).
The Turing test was developed by computer science expert Alan Turing in 1950. It assesses the intelligence of a computer by seeing if a human is able to tell whether they are interacting with a machine or a person. If the user can’t tell, the computer passes the test.
Panellists in a session discussing how bots are becoming a new interface between humans and machines gave various estimates regarding the timeline for bots to pass the test.
Vasco Pedro, co-founder and CEO of translation service provider Unbabel, said the fact bots are usually written in one language will hinder deployment and mean the Turing test won’t be passed for another ten years.
However, David Hanson, founder of robot maker Hanson Robotics, predicted the test could be passed in three years, because doing so depends more on “good craft” than advances in artificial intelligence (AI). In Hanson’s view, a well-authored chat bot can be an “interactive work of literature”, which is equally effective as a good novel in convincing readers a character is real
Alexander Del Toro Barba (pictured, left), head of product and marketing at Fintech company VisualVest, took the middle ground of five years. He told delegates he is impressed with technological advancements in China, where voice recognition is strong thanks to work done by Baidu.
Bots and their challenges
Adelyn Zhou (pictured, right), co-founder and CMO at bot search company Topbots, explained there are three main types of bot: customer service bots such as one covering check in and boarding passes used by KLM Royal Dutch Airlines; marketing bots as used by cosmetics retailer Sephora; and feature bots, which can be used to conduct mundane tasks normally handled by a person.
Customer service bots can enable businesses to scale up their CRM without increasing customer service costs, while marketing bots are “the new wave of social media,” Zhou said.
Bots can also be segmented into those based on text or messaging, and voice-based. Text-based bots offer a very basic interface which makes some features like shopping and product discovery difficult.
While most people are more comfortable using a voice-based bot, there are some drawbacks. For example, users may struggle to phrase their query correctly to elicit the desired response from their bot.
More worryingly, Del Toro Barba said a survey of 12,000 people found half of them believe they could fall in love with their virtual assistant.
The risk of love-struck users is worthwhile, Del Toro Barba argued, explaining virtual assistants and bots offer a “whole different level of customer relationship management” to industries including the financial services sector.