Bot chat – how writers are helping machines to talk

Following his recent CR article on copywriting and social media, we talk to Nick Asbury about writing, technology and @botconference, his Twitter bot that posts live updates from a non-existent conference – and poses questions about how language works in the process

Copywriter Nick Asbury recently launched @botconference, a bot that tweets imagined soundbites from an imagined conference, some of which are just a whisker away from exactly the kind of thing you might expect to see on an event feed. ‘Design thinking is the new authenticity,’ for example. ‘It’s not about social. TV is dead’. And RT and RT.

On the surface, @botconference is a perceptive parody account poking fun at the kind of conference platitudes aired on social media, but it also raises interesting questions about the way in which automation is increasingly playing a role in how we interact with brands – and what the writer’s role might be within this.

Asbury’s interest in the point at which writing and technology meet culminated recently in an article for CR’s ‘social’ issue that we published last month (you can read it here). As a follow up, we wanted to revisit the subject and talk to him specifically about his fascination with bots, LiveChat, machines and language.

Botconference-6
From @botconference

Creative Review: How did the idea for @botconference come about?

Nick Asbury: I came across a site called cheapbotsdonequick.com via Russell Davies and his @taglin3r tweetbot, which generates brand taglines. There’s been a lot of chatter about chatbots recently, so I was interested in trying it out.

CR: Why did you use the model of the conference for the experiment? They’re the perfect source for soundbites, which are perfect for Twitter – it does have a nice circularity to it.

NA: Yes, conference tweets are strange things. On one level, @botconference is a parody of the worst tweets – people peddling clichés about how everything is dead and the future will make no sense unless you buy into whatever new buzzword they’ve just invented.

Botconference1
From @botconference

But there’s also this weird phenomenon where even interesting speeches have a different effect when condensed to soundbites. Removed from their context, they always sound over-confident and over-prescriptive. It’s a good argument for going to actual conferences, because the real thing is usually better than the tweets.

But given that conference tweets already feel removed from real conferences, it’s not a great leap to remove the conference element entirely. There’s something poetic about the idea of bots talking to each other in an empty room – and it’s interesting to wonder if you can generate useful insights without the need for active thought.

CR: Have you noticed if many of the more believable tweets are getting any traction on Twitter?

NA: Some are getting retweeted by people with generic tech/thought-leader profiles. I suspect they’re automated RTs triggered by keywords – bots retweeting bots.

Taglin3r-1
From @taglin3r

CR: It’s interesting how @botconference occasionally produces a seemingly-insightful phrase, in the same way that Russell’s experiment also has some believable results. With his work in mind, can you tell me more a bit about this idea of taglines disregarding the rules of language that humans are prone to follow?

NA: @taglin3r is interesting, because it’s part-parody and part useful tool. I’ve written about the trend for brand lines that turn nouns and verbs into adjectives, or vice versa (Be Legacy, Be Your Way, The Do Inside etc) – and I conducted a jokey exercise of slicing them up and rearranging them into alternative lines. @taglin3r is the automated version of that cut-up technique. It’s produced some good lines – recent favourites include ‘We Speak Greatest’ and ‘For A World Of Cheaper Glory’.

If you want to make a trendily ungrammatical tagline, then bots are fast at generating possibilities, and also potentially more effective. As a trained writer, it takes effort to force your mind off the usual syntactical tracks. Computers don’t have that problem. @taglin3r turns the clumsiness of bots into an asset.

Taglin3r-2
From @taglin3r

CR: How do you set the bot up on cheapbotsdonequick – do you feed it key words and phrases and then ask it to publish them to a schedule?

NA: Yes, you set up a formula like ‘firstnoun’ is the new ‘secondnoun’. Then you enter as many values as you like for firstnoun and secondnoun and the software generates combinations. You can set it to tweet at regular intervals. It’s a nice interface, designed by George Buckenham.

CR: The CR piece you wrote touched on “the intersection of writing and technology, where interesting things are happening”. What kinds of things do you think writers might be involved with over the next few years? And as we hear more about the ‘internet of things’ – and giving personality and emotion to objects that we interact with – where does writing fit into that?

NA: There’s already lots of demand for writers to help machines talk. Siri was the milestone project – it’s easy to forget how successful that was. Apple could have created a straight Apple-branded voice assistant, but they created a character – a machine with a sense of humour and sometimes the appearance of self-awareness. I wish Douglas Adams had been around to see it – the Hitchhiker’s Guide was the original Siri.

Now the makers of Siri are developing Viv, an AI assistant that doesn’t just look things up online, but also gets things done for you – a kind of conversational user interface for everything on the internet. And Facebook is exploring similar territory with Messenger chatbots. There’s a lot of talk about ‘conversational commerce’ – talking to bots instead of typing into browsers or apps.

Viv-Dag-Kittlaus-presentation2
Still from Beyond Siri: The World Premiere of Viv with Dag Kittlaus, a TechCrunch Disrupt NY presentation on YouTube (May 9)
Viv-Dag-Kittlaus-presentation3
Still from Beyond Siri: The World Premiere of Viv with Dag Kittlaus, a TechCrunch Disrupt NY presentation on YouTube (May 9)

All those interactions will be scripted by writers, but it’s worth saying they won’t necessarily be copywriters. The developers creating this stuff are looking for poets, novelists, comedians and especially screenplay writers, because they are seen to have a better grasp of how to create characters with distinct voices and back stories.

Maybe the perception is that copywriters will write a version of marketing-speak. It’s not a fair perception because copywriters have many skills to bring – organising information, writing concisely, using wit and ideas. But it’s probably more important now for copywriters to be interested in other forms of writing too.

CR: As a writer, where does your interest in chatbots come from? 

As a copywriter, I’m struck by how brands have always been like chatbots, in the sense that they are weird, robotic things that try to be human. And people have always been suspicious of brands because at some level you know it’s all a charade. But if bots become the new interface between us and brands, that could be liberating. It introduces a new honesty into the equation.

Rather than pretending, a vague, ethereal thing like a brand can have a consistent ‘tone of voice’, you can have a bot that is an actual character. Imagine if every transaction with the Virgin brand was through a bot called Ginny who, as well as doing everything for you, was also clearly in love with Richard Branson and afraid to admit it to herself. It might be quite fun.

Or the bots could become like SatNav, voiced by different celebrities. You could order pizza by talking to Bob Dylan. God, this sounds terrible.

CR: And what is it about some of these experiences that people feel uncomfortable, do you think? Are we still suspicious of automation, particularly when it comes to the human ‘voice’?

NA: I think people are suspicious of anything fake or manipulative. If chatbots are used as a gimmick, or as a way to save costs and make life more difficult for the customer, then people will resent them. But Siri wasn’t like that – it wasn’t replacing anyone’s job, but offering something new.

Screen Shot 2016-05-11 at 15.37.37
From Pckwck

CR: You also mentioned things like LiveChat and new forms of writing. Can you say more about that?

NA: I’m intrigued by those LiveChats you have with service providers on websites, where you’re typing messages to a real person in real time. It feels weirdly intimate, but also less personal than having to pick up the phone or go into a shop.

Lots of people don’t like talking – given the choice, we send texts and emails. Talking is actually a weird form of communication – making grunting sounds with your throat – but I guess it’s about to make a comeback with conversational UI.

Related to LiveChat, there’s the kind of real-time collaborative writing you see in Google Docs, where you see the text appearing as people type. People have used that technology for experimental projects like writing a live novel (pckwck.com), but it’s never really found a wider use. I’m interested in that middle ground between conversation and ‘proper’ writing.

CR: You ended that CR piece with ‘What happens next might actually amaze you.’ Are you optimistic about the future?’

NA: I’m pessimistic about some aspects. For example, in the next few years, I think we’ll see a lot of brands launching superficial AI and chatbot projects mainly as a PR gimmick, and the media will lap it up. Chatbot will be the new ‘emoji’ in the sense of being a trigger word that the media automatically finds interesting.

I also worry that conversational UI will lead to a rise in talking over writing. And I worry what brands will do with it. We could move from the frying pan of chatty copy into the fire of actual chat. It could be grim.

But I’m optimistic about the bigger picture, like new forms of writing emerging, and writers working collaboratively with machines, and brands doing more interesting things with words. That’s an interesting time to live in.

Nick Asbury is a writer for branding and design and one half of creative partnership, Asbury & Asbury. See @asburyandasbury. his article, Too Long; Didn’t Read appears in the May issue of Creative Review, out now. 

More from CR

Highlights from the expanded edition of In Loving Memory of Work

Craig Oldham’s In Loving Memory of Work: A Visual Record of the Miners’ Strike 1984-85 has been published in a second edition, complete with a range of new material. Here, Oldham selects some of the most interesting additions, including ex-miner Phil Winnard’s powerful photographs of the strikes.

D&AD Next Photographer winner Tam Hoi Ying on challenging censorship in China

Hong Kong-Chinese photographer Tam Hoi Ying has received this year’s D&AD Next Photographer Award, in partnership with Getty Images, for a series which raises awareness of issues surrounding freedom of speech and human rights in China. Here, she discusses her work and how she hopes to challenge censorship

Lecturer Design Management

Kingston University

Graphic Designer

Fishfinger

Design Assistant

Cultureshock Media