Create Creative Bots

Can bots help us be more creative? Can they help us produce art?

Absolutely. An important way to sustain human creativity and to magnify human creativity and to keep vibrant communities of human creators going is to embrace more, not less, automation, and to embrace the idea that bots themselves can be creators that work with us.

It’s hard to be creative on schedule. So for a very long time, psychologists and creative practitioners have studied ways to systematize creativity – to put creativity at our fingertips and to bring creativity when we need it.


The most popular and most widely known is brainstorming. You put a bunch of people in a room and you get them bouncing ideas off each other. Brainstorming works because it allows people to suggest ideas in a nonjudgmental context. It doesn’t actually deal with the process itself of creating ideas. To be truly creative and to systematize real creativity, we have to drill down into the structure of ideas and talk about how new ideas are formed, not how to treat them when they are formed.

This is a very common practice. Go to any cinema showing a Hollywood movie and you will see the results. This is a classic Western set in space or at sea or on top of a mountain. Hollywood does this instinctively. It’s called the theory of inventive problem solving and again both of these procedures allow you to be creative or help you to be creative by giving you lots of tables and charts. It’s a bit like playing Dungeons and Dragons actually or using a spreadsheet. This kind of thing works.  

Twitter bots aren’t new

The twentieth century was the most abstract ever, but it was also the century in which art became the least. Artists looked for ways to least semi automate the production of new ideas. They were really interested in producing more and more and more are, but they were terrified of repeating themselves and they were terrified of succumbing unconsciously to cliché and so they came up with a whole bunch of tricks and rules and random techniques to generate new ideas.

William S. Burroughs really pioneered this technique. Burroughs and others used newspapers to source their writing for the cut-up technique, and we can probably agree that today, Twitter is basically the newspaper.

In this method you recombine text, and you seek out the random combinations that have surprising meanings. It was really brought to the fore by David Bowie.

He is famous for using this technique to generate possible ideas for the lyrics in the songs, which is why the lyrics in his songs are so weird: “Ziggy Stardust and the Spiders from Mars.” Radiohead did the same thing with their album “Kid A.” These artists are trying to avoid cliché – we’re all slaves to cliché, that’s what it means to be human. Clichés work. That’s why they’re popular. But we want to avoid it when we’re being creative. So to combine ideas you have to do a lot of filtering to take the ones that work.

Picasso famously said that computers are useless – they can only give you answers. Google is a great at giving you answers. It’s not so good at telling you what to do next. If we want bots to be creative they have to pose interesting questions, because good art doesn’t provide an answer. It asks a question.


This field I’m talking about is called computational creativity. It’s a branch of AI but it’s also the stepson of modern art because in computer science we’re always using ideas that were invented somewhere else. So creativity you can think it as the next step – but we have to be careful if we train our bots on convention and to recognize when something is not conventional then it won’t recognize true creativity or it will treat true creativity as an error. We need to stop focusing on the surface and we need to start focusing on the inner workings of human creativity. We need to study humans and see how humans are creative and to transfer that cognitive working into algorithms in a bot.

We can understand how humans are creative in algorithm terms, which can help humans be more creative as well. We want creativity in a bot to expand human horizons to ask the questions that humans don’t ask because we’re enculturated not ask them or we’re too polite, or because we don’t consider them as options or because we can’t see ourselves from the outside. Bots can do all of these things. Treat them like an anthropologist from another planet. A good bot program can look at human behavior and see the silliness and or see the interesting parallels between different facets of human behavior and report on them and make interesting analogies, Ironic observations, and look at us from the outside in.

It only happens when we treat bots not as people but as creative authors. We want our bots to compliment. Creativity often arises from the conflict between people with different agendas and different goals. So we want our bots to be different from us so that we get this conflict that sparks new ideas.

The myth of the lone genius has held back creativity studies for a long time. Genius comes from social interaction. No matter how lonely the creator, he or she has benefited from interacting with others. So we need to make our bots social beings. Not humans, but social beings that know how to interact with humans and benefit from the interaction. In many ways a bot is pretty dumb, and we can use that dumbness to our advantage when it comes to making the bot help us be creative.

Where Twitter Bots Come In


The city is a great mentor for creativity. You push people of different backgrounds and agendas together. We can’t get our bots to be creative sitting on a laptop or desktop in the office. We need to get them into the city of course I’m not talking about putting them into robot bodies.

I’m talking about virtual cities. I think Twitter is a great virtual city for our bots to live in.

If you tweet often you’ve probably had some random bot “@” you for using the wrong “your”, or being a communist. While this might have just annoyed you, we can actually learn a lot of lessons about the future of bots, and the future of art.

Twitter bots come in many shapes and sizes. These bots are programs that are coded by humans and then sent to the cloud. The boring ones will just tweet a link at you about ipads. Some you can buy in order to pump up your followers. But the really cool ones take advantage of Twitter to create something truly unique.

Everything happens so much

— Horse ebooks (@Horse_ebooks) June 28, 2012

@horse_ebooks is my personal favorite. It was maybe the first account to make people think that Twitter was something special – or something very, very weird. @horse_books at least initially was a legit spam bot designed by a Russian programmer along with about one hundred seventy other bots to sell e-books. What’s interesting about horse_ebooks is that it was terrible at its job.

Its sole purpose is to get you to buy these ebooks about adopting a horse – but no one does that.

The result was infamously, accidentally poetic.

Avoid situations

— Horse ebooks (@Horse_ebooks) July 17, 2013


The first deliberate use of code in Twitter’s open API was @everyword, which first tweeted in 2007.


— everyword (@everyword) June 7, 2014

@everyword was the dawn of a new Twitter bot era – one designed to make a deliberate statement every thirty minutes. After this, you see stuff like @betelgeuse_3 that’s always listening and replies when someone says “Beetlejuice Beetlejuice Beetlejuice”; or @big_ben_clock, that tweets every hour; and many, many more.

The X-Files
Oct 1996

— wayback_exe (@wayback_exe) December 10, 2015


Are Millennials Getting High On Beanie Babies?

— Thinkpiece Bot (@thinkpiecebot) November 13, 2015

Of course the ability to pull information from tweets goes beyond usernames.

Here’s an example of a wonderful bought called @pentametron. It creates poetry by taking tweets of other people and pairing them up. It won’t meet formal constraints of poetry.

my hearts already breaking twist the knife .

— britney (@versacebret) August 17, 2016

Appreciate the little things in life

— J.T. Young (@jyoung538) August 17, 2016

It’s using Bill Burroughs cut-up technique in a very clever way by combining two tweets to make a couplet. Imagine what we could do if we added power into these simple techniques on Twitter where social interaction gets magnified. It’s very democratizing for creativity.

Twitter is a level playing field. If your bot or you are doing something good you attract followers. In fact a great deal of the intelligent content or the creative content on Twitter is produced by bots right now and people knowingly follow those accounts because they are sick of listening to humans.

Maybe you have heard the dire predictions from famous scientists saying that AI is the genie that will destroy your soul if we let it out of the bottle. When CERN built a large hadron collider, some people worried that it would create a black hole that would suck the whole world into nothing. The physicists laughed at us. bots will not be the ends of us. They’ll be the making of us because they will be co-creative partners that will extend what it means to be human.

And anyway, William S. Burroughs would have loved a computer program that randomly paired the subject from one newspaper headline with the action item from another headline for often hilarious results.

Twitter Was Robbed At Gunpoint In Brazil

— Two Headlines (@TwoHeadlines) August 15, 2016

Donald Trump 'yellowface' filter removed after being accused of racism

— Two Headlines (@TwoHeadlines) August 12, 2016