For the first time since 1947, the Edinburgh Fringe Festival as we know it is not going ahead. But there are years of previous programmes stored in the digital vaults. The Bot, an artificial neural network, goes to work on these data, compiling the world’s first AI-generated event blurbs for a virtual arts festival of comedy, plays, musicals, and cabaret… will you lend a hand in the effort?
The aim of ImprovBot is to explore the junction of human creativity and comedy, and to see how this is affected when an artificial intelligence enters into the mix. ImprovBot was originally conceived as a live show for the 2020 Edinburgh Festival Fringe, with Edinburgh University researchers collaborating with the world-renowned Improverts, the Edinburgh University Theatre Company’s resident improvised comedy troupe. We had planned to load the previous eight years of programming of the Fringe, which the Edinburgh Festival Fringe Society had made available to us in digital format, onto the UK’s National Supercomputer (at the University of Edinburgh). We would then use our neural network, The Bot, to generate “new” show blurbs from the 2.5 million words of stored data. Using a combination of live and pre-computed AI-generated results, we would challenge the Improverts to perform these new proposed shows on the fly, with presumably hilarious results. Our social media presence would also publicise these virtual shows, with illustrations that counterpoint the generated text, providing an extended meditation and critique of the mushrooming stock image industry, and clichéd depictions of AI, computer science, and academic research. We would see whether online audiences would rise to the provocation, and interact, remix, mashup, and play with the content we produce. We would then stroke our chins, to ask what this all meant.
Unfortunately, given that the 2020 Edinburgh Fringe is not going ahead, the live show cannot go ahead as planned. We have decided to still bring ImprovBot to you in Festival season 2020 as a reminder of the playfulness of the Fringe, to build on remembrances of Fringes past, and to bring some levity at a time when digital connection has become ever more important for us all, during the lockdown of 2020. What tropes and trends will emerge? Do the ghosts of festivals past affect festivals future? Will it make any sense at all? And how can we distantly collaborate with these improvisations in unprecedented times, during normal festival season? We encourage you to respond to the provocations set by the AI, and to interact with us digitally. We will all miss the Festival. The Bot is looking forward to your collaboration!
Our thoughts are with our colleagues in health and social care, and other workers supporting us all during the COVID19 pandemic. We are aware of the economic devastation caused by the cancellation of the Fringe for artists, writers, producers, reviewers, venues, backstage crew, and also the knock-on effects this will have for many throughout our home city of Edinburgh, including the Festival organisations themselves. We hope the festival community is safe, and healthy, and we at the University look forward to working with everyone in future. We are eager to bring ImprovBot: Live to the Edinburgh Fringe Festival in 2021.
And we look forward to your interactions with ImprovBot!
Access to the Edinburgh Festivals Listings API, and permission for creative (non-profit) reuse of previous programme information, was granted with kind permission from the Edinburgh Festival Fringe Society. We also thank Daisy Francis, Managing Producer of the Improverts, for being receptive to our developing ideas, and the Improverts for being keen to experiment with ImprovBot.
This project has received ethical approval from the University of Edinburgh’s School of Literatures, Languages and Cultures.
In a pre-Covid19 world, ImprovBot received small project funding from the Edinburgh Futures Institute, in conjunction with the Data Driven Innovation program of the Edinburgh and South East Scotland City Region Deal, with additional resources from the Edinburgh Centre for Data, Culture and Society, and Creative Informatics. We are also grateful to our colleagues at EPCC for their provision of access to computational resources.