A Brief History of Chatbots

It’s hard to think of a more cutting-edge solution than a chatbot. Given the advanced technology that powers today’s versions, most of us see these digital helpers as a truly modern concept. But in actual fact, it’s now over fifty years since the first mainstream chatbot blew our minds. 

Since then, the technology available has changed the mechanics behind chatbots, but the goal has always been the same – to create machines that we can interact with naturally, via speech or text.

To help us get a better understanding of how we got to this point, here are some of the landmarks in the chatbot’s long and winding history.  

1920s – the rise of ‘robot’

Long before we had the technology to realise the dream, we philosophised about the idea of intelligent machines, particularly in popular science-fiction literature and film. In fact, the word robot was first coined by Czech author Karel Capek, and comes from the Czech word robota, which means ‘forced labour.’

1950 – the Turing Test

Several decades later, almost 70 years ago, Alan Turing, the legendary British mathematician and computer scientist, devised a protocol for establishing whether a machine can truly be thought of as intelligent or not – known as the Turing Test.

The Turing Test was simple enough: if a human cannot distinguish between another human and a machine in a conversation, then the machine can be considered intelligent. The Turing Test has since become an important concept in the field of artificial intelligence (AI), and laid the foundations for the development of modern chatbots.

1966 - Eliza

Created in the research laboratories of Massachusetts Institute of Technology, Eliza was the world’s first high-profile chatbot. Designed to mimic a human therapist, Eliza was surprisingly effective at providing relevant responses to its ‘patients’ input.

Eliza worked by recognising key words, and then selecting pre-prepared question responses based on them. For example, if the patient typed ‘I feel exhausted’, Eliza might reply ‘Why do you think you feel exhausted?’, and so on.

Although this gave the illusion of intelligence, Eliza’s conversational ability involved a purely mechanical matching process; the program was unable to learn through experience, and had zero understanding of context.

1970s – Interactive Voice Response (IVR)

If you’ve ever contacted customer services only to be greeted with an automated voice response, you’ll know what IVR is. These much-maligned systems took off in the 1970s, and are still widely used today.

While not strictly chatbots, IVR is a good example of early attempts to automate conversations, replacing the need for a human operative when directing customers to the right department, or even handling simple processes entirely.

Like Eliza, there is no intelligence behind IVR. The system merely presents the caller with a range of options, and asks them to enter their choice using a keypad. While this system worked insofar as it allowed businesses to become more productive, the user experience was, and still is, famously frustrating.

1972 – Parry

On the other side of the couch from Eliza, Parry was a chatbot designed to mimic a paranoid schizophrenic. Parry took the art of conversation a notch further than Eliza; rather than submitting responses based purely on keywords, it followed a complex conversational strategy, which involved consistently misinterpreting the motives of others.

In a version of the Turing Test, a group of psychiatrists was only able to tell the difference between Parry and real schizophrenic patients 48% of the time.

1988 – Jabberwhacky

Designed to simulate human conversation in an entertaining and humorous manner, Jabberwhacky stood out from its predecessors as the first chatbot with the aim to move away from text-based dialogue and towards a voice-operated system.

Also, unlike earlier incarnations, Jabberwhacky made use of short-term memory – i.e. it retained some of the information from the previous input, and then used it in constructing its response. Although Jabberwhacky didn’t understand the meaning of the information it was retaining, this process brought a degree of shared context to the conversation – a step closer to what takes place in interactions between humans.

1995 – Alice

Standing for Artificial Linguistic Internet Computer Entity, Alice has won the Loebner Prize – an award given to the most human-like artificial intelligence – three times. However, despite its successes, Alice has failed to pass the Turing Test.

Despite its drawbacks, Alice was a ground-breaking effort at the time, and impressed users with its ability to answer pretty much any question. It did this using a technique called linguistic deflection, whereby the system produces answers to questions it doesn’t understand by deflecting them. For example, if you ask it ‘why are my socks blue?’, it replies ‘the answer is very complicated.’ This technique makes it very difficult to catch Alice out, even with impossibly abstract inputs.

2001 – SmarterChild

Considered a precursor to modern digital helpers such as Siri, SmarterChild was released on AOL and MSN messenger platforms, reaching a ready-made audience of millions. SmarterChild differed from previous chatbots in that it was designed to be more than just a novelty.

By connecting it to online databases, SmarterChild could respond to questions with real-time information. Whether you wanted to find out the football results or film times, it could locate the data and relay it to you. This was a significant step in the evolution of chatbots – for the first time, they had real utility for the end user.

2010 – the rise of modern digital assistants

Over the last decade, chatbots have developed from an online novelty to an everyday tool. Back in 2010, Apple launched Siri, the first widely available personal digital assistant. Competing models followed from tech giants Google, Amazon, and Microsoft.

These digital assistants use natural language processing (NLP) to understand the nuances of human speech. Hooked up not only to the internet, but also our personal diaries, libraries and contact lists, they are able to perform countless tasks on our behalf. Now that the initial stage of mistrust is passing, we can expect to see more and more people augmenting their lives with digital assistants.

2016 – Bots for Facebook Messenger

In recent years, chatbots have expanded the scope of their use beyond personal assistants and into the commercial realm. This chatbot revolution has coincided with an explosion in messenger app use, which surpassed that of social media in 2015.

Unsurprisingly, Facebook Messenger has since become the biggest chatbot platform, with over 300,000 active messenger bots as of May 2018. This has revolutionised the way customers interact with brands online, providing instant 24/7 access to customer support, automated e-commerce, and an entirely new approach to online marketing.

The future and beyond

Today’s chatbots may be 50 years in the making, but the chatbot story is far from over. In fact, this is really just the beginning.

The coming years will see chatbot use in businesses become the norm for both internal and external communications. Research shows that people increasingly prefer to speak to a chatbot than with a real human in areas such as customer services, and as adoption increases, our preference for chatbots will only grow. In just a few years, chatbots could become our go-to platform for all our technological needs, replacing the need for many apps and websites.

In many ways, the chatbots we have today have already achieved what we once dreamed about – to create intelligent machines that we can interact with, and which serve to make our lives easier. But there is still much progress to be made. Technologies such as artificial intelligence and natural language processing are a work in progress. As we make greater strides in these areas, chatbots will likely become an indispensable part of our lives.


Nicholas Edwards - http://www.praguecopywriter.com

Nicholas Edwards is a freelance writer and editor based in Prague, the Czech Republic. When he's not helping local businesses master the English language, he loves writing about the future of work for People First.

Linkedin Logo