BrainyBeard | An Investigation into the chatbot, is it relevant yet? - BrainyBeard

An Investigation into the chatbot, is it relevant yet?

Posted by  on 
30th March 2017

Does the Chatbot eliminates the need to talk to real people anymore?…

What is a Chatbot?

A chatbot is a computer program designed to simulate conversation with humans. Chatbots usually belong to one of two categories, the ‘utility chatbot’ or the ‘content-driven chatbot’. Utility bots helps the user achieve a task, for example booking a meeting. Content-driven bots provide some type of information, such as the latest news.

The key benefit of chatbots is that they allow for natural communication between humans and machines. This software works by either learning to communicate better using machine learning, or through static rule based systems. We are currently interacting with these bots mainly via messenger applications on our phones. In the last year, the use of bots has started to revolutionise the interface we use to access the internet.

Why Do We Need Them?

The purpose of technology has always been to make our lives better. Cars make us move faster, medical machinery helps us live longer and tools make our lives easier. The chatbot is a tool that resolves the things we aren’t very good at.  Finding clothes online that suit our style, correctly filling out our tax returns and remembering to do things are a number of tasks that humans struggle with.

What if this was your typical morning?

Your alarm goes off at 7am, which tells the coffee machine downstairs to start making some coffee. Meanwhile your fridge has determined it’s out of marmalade. It asks your assistant bot to add “Buy more marmalade” to today’s to-do list. At 7:30 your travel bot has warned you there are delays on the rail service. It asks you if you’d prefer it to order you a taxi?

As you can see, this technology is exciting! But how did it get to this point?

Eliza the original chatbot simulated a Rogerian psychotherapist…

The Original Chatbot

Chatbots are nothing new. In fact, the original chatbot Eliza was built by Joseph Wiezenbaum back in 1966. Only 16 years after Turing had written an influential paper that introduced the Turing Test. Turing’s Test is, simply put, meant to determine whether or not a machine can show intelligent behaviour in a way that passes as human. Although basic, Eliza simulated a Rogerian psychotherapist helping users better understand their problems. This proved chatbots had potential. VentureBeat produced this interesting infographic detailing the history of chatbots from Eliza to the modern day.

One of the most famous recent examples of a chatbot is SmarterChild developed by ActiveBuddy for AIM and MSN. Prior to being shut down, it provided users with information from the web. It also took a lot of abuse from angsty teens. Recently we have seen the rise of Twitterbots. These bots are aimed at RT’ing tweets with certain hashtags, or replying to tweets that contain specific content. None of these have seamless integration within our daily lives to help us communicate with machines and boost productivity.

What we have now?

Most of us still don’t use chatbots, and this creates an atmosphere of wonder around them. Combine this with our current fascination of artificial intelligence and it is no surprise that chatbots are in high demand. Unlike in SmarterChild’s era, there is now greater accessibility to tools to quickly create and test bots. This is also fuelled by developments in artificial intelligence, by companies like Google. Using Messenger, you can already talk to bots, go try it!

2017 could be the year we see ourselves taking a step forward to increasing our reliance on chatbots. Restaurants including Pizza Hut, Dominos and Taco Bell are already offering the ability to order using chatbots. Dominos allow this by simply tweeting a pizza emoji! Slack has Slackbot, which allows you to set reminders for yourself or members of your team built into the platform. Need info about your flight? There’s a bot for that. What about a quick quiz whilst waiting for the bus? Yep, bots have you covered.

Unfortunately, as the technology is still new, it does have teething problems. Language processing is still clunky and can create frustrating experiences for the user. They are often slower to use than simply doing the task yourself. It’s also inevitable that as bots take on more responsibility, we must ensure that they are reliable. For example, how often will it order the wrong pizza? AI can often act unpredictably so this could be a concern. Currently if the bots do make mistakes, they try to appear more personalised and human such as using emojis. Yet like all technology, if it doesn’t work it’s always going to be infuriating.

Where we are going?

One of the biggest questions surrounding bots is how will they be monetised? Naturally, we would expect the underlining technology to improve, but could it come at a premium? Will we have to pay to use them, or does the company save money by removing their customer services team? Bots would have to be more sophisticated and the expectation would be that the user never interacts with different bots. Instead, we would have a personal assistant bot that acts as the middleman.

What is clear however, is the enormous potential bots have in a commercial and educational setting. Imagine walking into a shop and a chatbot welcoming you on your phone. A shop bot could potentially point you to any deals suited to your tastes. In education, it’s easy to think about how you could go paperless with your diary. Imagine asking your bot to record any homework that has been set or informing you what class is next.

As with all new technology, there is hype that over-exaggerates the current technologies capabilities. Chatbots will only improve drastically when the market is ready to adopt them on mass. Virtual reality is a good example of this. We were supposed to adopt the headset in 2016, yet the majority of us haven’t even tried the Google cardboard. The more powerful (and higher priced) HTC Vive and Oculus Rift also seriously lack any great games.

It seems communication with other humans isn’t dead just yet…

Should We Be Concerned About The Chatbot?

Will chatbots help robots take over the world within the next 30 years? Probably not. The biggest challenge we’re likely to face in the coming years is the ethics behind using chatbots. Is it a problem if you don’t know whether you are talking to a human or not? Must we adopt some kind of standard that makes it very clear when you are engaging with a bot? Furthermore, there are concerns involving the psychological impact that developing companionship with machines has on the brain. Her, a 2013 film, has an interesting take on this subject. These are issues that as a society we have been slow to react to and will become a problem.

Chatbots are fascinating and to think about how the technology will integrate with our lives is exciting. What are your thoughts about chatbots? Have you used them? Let us know your thoughts over at Twitter or Facebook.