Harry Potter and the Prisoner of… Algorithms?

Last week, the Internet went nuts over a new chapter of the Harry Potter series… that was generated by a predictive text algorithm programmed to generate text in the style of J. K. Rowling.

‘Harry Potter and the Portrait of What Looked Like a Large Pile of Ash’ was generated by Botnik Studios, an entertainment group that displays the work of a self-described ‘community of writers, artists and developers collaborating with machines to create strange new things’. Botnik artists have produced things like a new Scrubs script, a computer-generated History of Thanksgiving, and some pretty wacky reviews for the latest Star Wars film. Yup, what they do is strange, and their latest edition to Harry Potter is no different.

‘The castle grounds snarled with a wave of magically magnified wind’, the chapter begins. Seems innocent enough.

Next sentence: ‘The sky outside was a great black ceiling, which was full of blood.’ Huh. That’s a strange image.

Moving on: ‘The only sounds drifting from Hagrid’s hut were the disdainful shrieks of his own furniture. Magic: it was something that Harry Potter thought was very good.’ Good God, Hagrid’s in danger of his loveseat and Harry’s become a psychopath.

The text continues in such a way. Here are some of my favourite passages:

  • ‘Ron was standing there and doing a kind of frenzied tap dance. He saw Harry and immediately began to eat Hermione’s family.’
  • ‘Ron was going to be spiders. He just was. He wasn’t proud of that, but it was going to be hard to not have spiders all over his body after all is said and done.’
  • ‘The pig of Hufflepuff pulsed like a large bullfrog. Dumbledore smiled at it, and placed his hand on its head: “You are Hagrid now.”‘

The creators are so proud of their work that they’ve even had the story printed and made to look as through it were a legitimate addition to the series.

Someone else has taken it upon herself to illustrate some of the funnier passages.

NME deems the output ‘better than anything in Morrissey’s novel.’ The Guardian calls it ‘gloriously bonkers’. Mashable says it’s ‘wild’, but ‘let’s just say we’re glad it’s not canon.’ When one YouTuber posts a video of herself reading the story aloud and providing her own commentary, the comments section is overwhelmingly positive.

As many followers of my blog know, my doctoral research is all about the social and literary implications of computer-generated texts. I spend about seven hours a day, at least five days a week, sitting at my computer and thinking really hard about this kind of thing. I run focus groups and online surveys (take my reader response survey if you haven’t already!) to get to the bottom of where computer-generated text actually fit within our current cultural landscape.

In short, this kind of thing is right up my alley.

And this ‘Harry Potter and the Portrait of What Looked Like a Large Pile of Ash’ thing just doesn’t seem to fit with the data I’ve collected.

I’m still very much in the analysis stage of my research. Nothing about the qualitative data I’ve collected from my focus groups and reader response survey has been conclusive, and I don’t want to give anyone the impression that I think of myself as some sort of media prophet when it comes to natural language generation. But the Internet’s responses to this computer-generated text has been noticeably different from the hundreds of responses I’ve gotten, and I think it’s worth taking a moment to think about why that may be.

I’m going to go ahead and predict that these contrasting responses are mainly due to the entire world’s undying love for anything Harry Potter, as well as the novelty factor of computer-generated texts that blinds readers to conducting critical analyses of the text’s method of production. Not to say that there’s anything wrong with the text’s method of production – the predictive keyboard supposedly used to generated this text is similar to that found in mobile phones’ text messaging applications (although we’re not sure how much human intervention the text has seen). However, my own research has shown that people tend to get pretty uncomfortable once they’ve been told that a text they’ve been given has been computer-generated. My focus group participants have indicated that they’ve felt deceived. And then there’s ‘the robbery of the human soul’. Don’t even get me started on the human soul. What does it mean?!

It’s clear that Botnik isn’t out to replace the beloved J. K. Rowling any day soon. It’s clear that ‘Harry Potter and the Portrait of What Looked Like a Large Pile of Ash’ was produced for innocent entertainment purposes, and wasn’t intended to demonstrate how human authors will one day be altogether unnecessary. Readers seem to understand this, and therefore approach the text with open, rather than skeptical, minds. They want to laugh, and they want Harry Potter. ‘What Looked Like a Large Pile of Ash’ gives them an opportunity to engage with a world with which they are familiar, while at the same time allowing them to feel like they can chuckle off any potential ‘Singularity‘ thoughts they may have had. If anything, it’s a great entry into algorithmic authorship: people need to know that this stuff is happening before they can think about it in any depth.

This is just a short blog post to let readers know that I’m aware of this Harry Potter thing, and to get people to stop sending me news articles about it (just kidding!). My research delves much deeper into this kind of stuff, and presents actual findings and theoretical considerations much more eloquently, so keep on following me for more. I may come back to ‘What Looked Like a Pile of Ash’. I may not.

I’m not really a Harry Potter fan. Sorry.

New Directions, Networking… and Cookies

Well, folks, my first conference presentation as a PhD student was a success!

This past Tuesday, I had the pleasure of presenting some of my research at Oxford Brookes University’s annual ‘New Directions in Print Culture Studies’ symposium, hosted by The Oxford International Institute for Publishing Studies. This conference featured the wonderful Drs Shafquat Towheed and Samantha Rayner as keynotes, as well as a bunch of other academics investigating a vast range of topics within the fields of book history, print culture, and publishing.

My research is still very much in its early stages, but those who heard me speak were so supportive of what I was doing. I didn’t scare anyone away from natural language generation!

Want to give my presentation a peruse? Here you go:

One of the best parts of any conference is meeting people within your field who can challenge your views, but who can also offer you the support you need to continue improving as a researcher. As with every conference, I did do some pretty typical networking, and ended up meeting some interesting people from across the UK and Europe.

Here’s a hot tip: If you’re not great at starting conversations with strangers, just stand ever so slightly in the way of wherever the cookies (or other sweet treats) are. When someone realizes that he/she has to speak to you in order to get you to physically move, that is when you make your conversational move. Ask that person what he/she has found to be the most interesting part of the conference so far. Grab another cookie for yourself as you do so. Once both parties have a cookie in hand, y’all are practically guaranteed a jolly chat. You’re welcome.

Seriously, though, while networking went well, it was the panel I was assigned to that made my day. I sat on a panel with two other women – Dr Kate Macdonald from the University of Reading, and Laura Dietz from Anglia Ruskin – whose research topics differed drastically from my own. However, as we got to talking and presenting we realized just how much overlap there was between what we were saying. Following our presentations, Kate offered me some ideas about potential research avenues (which I’ve already begun diving into), and Laura provided insight into some ways to collect qualitative and quantitative reading response data (which will be very helpful for next year).

Also, we now have this photo where I look like I am going to eat Laura:

Sorry about that, Laura.

Hopefully I’ll be attending more conferences soon, although for now it’s back to the grind of chapter writing.

Got questions about the presentation embedded above? Hit me with ’em!