What’s a Chatbot (And Why Do I Need to Know)? Part 2 by Oisín McGann
I write science fiction, and one of the things about a passion for sci-fi is that it often stems from curiosity; you’re driven to exploring possibilities. These days, writing near-future science fiction is becoming more and more challenging, because things are moving so fast, that every time you try to anticipate what’s coming next, you discover it’s already happened. It’s hard to keep up. In real life too, technology is developing so quickly that we’re struggling to learn the skills to deal with it. When you get right down to it, it’s a language issue.
Publishing has always been about using technology to convey language to other people – but as it becomes more complex, we’re having to learn new languages to connect with the technology itself. We have so many new ways to communicate and yet these are additional skills we must learn; that communication seems to create new frustrations even as it solves old problems. If you’ve ever lost your temper with a computer (and I bet you have), or wondered if you can be bothered getting to grips with another new app, you’ll know what I’m talking about. It’s not their fault; they’re not trying to be dense, they just don’t have the capacity to understand you.
Writers, for instance, have a hard enough time communicating clearly to human readers, counting on them to be able to use imagination, to infer meaning from new constructions of words and to engage on an empathic level. Computers can’t even meet us halfway like this. Type some words, click a mouse, touch a screen, speak into a microphone. For you to be able to do this, you had to put valuable time and effort into learning how to understand the application – and someone had to figure out how to make the application understand you. Even for normal voice commands, someone had to programme the computer with a vocabulary of sounds it could refer to and give you instructions that told you about this vocabulary. Okay, so now you’re talking instead of typing, but it’s really only a replacement for point-and-click. It still can’t understand normal speech. You’re choosing from a limited menu of options and your voice commands have to be specific and very clear.
As our interaction with the web become increasingly complex, we’re faced with greater demands to learn how to use each new app we engage with. Chatbots are intended to cut down on the skills needed to use technology. They’re a means of making our computers more responsive, more personalised to our needs, in order to make our interaction with it feel more natural. And actually, the idea of a chatbot isn’t even that new. In 1966, Joseph Weizenbaum, at the MIT Artificial Intelligence Laboratory, created the first chatbot, a programme that could mimic human conversation using pre-prepared, scripted responses to key words. He wrote dialogue for it. He was trying to prove how superficial communication was between humans and machines. Instead, a lot of people found themselves charmed by what, for the first time, seemed to be a friendly computer. Despite Weizenbaum’s frustrated assurances to the contrary, some people even thought his programme, ELIZA, was intelligent, and understood what they were saying. Writers dream of creating this kind of magic; the scientists were just bemused.
Chatbots Are Not Human
A modern chatbot is not intended to be a cheap electronic copy of a human. It is a smarter, more versatile and sophisticated form of the kinds of apps you use every day – or it can use your apps for you, an assistant in your computer, doing the work that bots do so well; searching, collating, analysing and refining information, and it can do these things a lot faster than you can. It can compare the costs of flights or insurance quotes, find a book online or suggest one based on you’re the books it knows you’ve read . . . and you can ask it the way you’d ask a human and it will understand the request. Technology like this reduces the skill and time you have to invest in any action, making the interaction between humans and computers more intuitive, to make it feel more natural.
So Where’s This All Going?
Today, Apple’s Siri, Microsoft’s Cortana and Amazon’s Alexa are all examples of a sophisticated chatbot known as an ‘intelligent personal assistant’. These companies are investing heavily in this technology. Chatbots’ human-like responses are just the logical step up from the touchscreen, which grew from the point and click of a mouse, which in turn grew from typing code onto a screen. Each new step demands less skill, and takes less effort.
These examples are the top end of chatbots, created by huge tech companies on the inexorable climb towards true artificial intelligence, dreaming of the likes of Jarvis in Iron Man. The simpler end of the scale, however, is fast becoming more cheaper, more practical and more versatile, just using type instead of voice. Bot stores are expected to take over from app stores. For an industry like ours, where we trade in words themselves, this technology offers a dizzying array of pitfalls, but also tremendous possibilities.
Even as we find ourselves competing for readers with social media, we are offered an increasing number of ways of connecting to them. A friend of mine, the writer Conor Kostick, is creating a feature on his Facebook page using Snatchbot on his messaging app to offer sci-fi and fantasy book recommendations based on how a visitor answers a chatbot’s questions. If you’re a publisher, imagine your website being able to answer visitor’s questions live online. Bots have been used to perform as characters in games for a long time, but imagine if a reader was able to have a conversation with one of their favourite characters.
For years, we’ve used spellchecker and other functions to refine our writing; Scrivener offers a popular means of organising the elements of any written piece . . . now imagine an assistant that did all these of things, while being able to take dictation, record notes on the fly through your phone, compile them with your work file, or even have conversations with you that might prompt ideas. The same assistant could soon be able to help you spot interesting book events, or organise the bookings for your own events: timing your talk, checking your diary, bus or train timetables, flight times, hotel vacancies, seeing if friends in the area are available to meet up. It could track royalty payments, remember payment dates, production schedules and remind you about that thing you wanted to add at the editing stage of the latest book.
These are all functions of existing apps; soon, they’ll be integrated into a more intuitive system where you’ll just ask your assistant to do something and it will find the app that does it, if it’s not already part of its makeup. And, of course, to make all this feel human, feel natural, all these scripts are going to need to be written. They’re going to need good dialogue. Because computers don’t do colloquialisms, or analogies, or wordplay, or humour. Character, an engaging voice, will become an important selling point.
Perhaps it’s ironic, but even as writers increasingly rely on computers to interact with the world . . . computers are going to need writers for exactly the same reason.
(c) Oisín McGann
Oisín McGann is a writer and illustrator who has produced dozens of books and short stories for all ages of reader and has performed thousands of children’s sessions in schools, libraries and festivals. He is the author of the Mad Grandad books, the Forbidden Files and novels such as Rat Runners, The Gods and Their Machines and The Wildenstern Saga. He served on the board of Children’s Books Ireland for seven years and is a winner of the European Science Fiction Society Award, the Bisto Book of the Year Merit Award and has been shortlisted for numerous other awards, including the Waterstones Childrens’ Book Prize in the UK, le Grand Prix de l’Imaginaire in France and Locus Magazine’s Best First Novel Award in the US.
Read Part 1 of Oisín’s article here.
Order Oisín’s books online here.