We had a fascinating day of presentations and discussion at the Art AI Festival’s one-day conference. This post features our artist speakers who all work with creative artificial intelligence (AI) but use it in different ways. The post includes short summary videos by each speaker and is intended to give you a flavour of the topics and perspectives presented about creative AI practice from the practitioner’s viewpoint.
Ernest Edmonds, creative artificial intelligence pioneer and Professor of Computational Art at the Institute of Creative Technologies, began the day’s proceedings with a keynote that set the scene for the presentations and panel discussion later in the day. In his talk, he set out to ask some important questions: can AI make art; can AI be the artist; or, can AI be helpful to the artist? He points out key differences between connectionist (eg., neural networked) and symbolic AI: the former has many issues currently associated with it, not least some major ethical concerns, whereas symbolic AI makes its description and application clear at the outset. He positions his work in the latter domain, where the AI is adding to creative possibilities, increasing possibilities for audience engagement and enables him to build more responsiveness into the artwork. Ernest states that AI has been a positive force in his creative focus over the last 20 years, stressing “… [my] artwork learns from the audience and evolves and develops in itself, over time, which matches how we evolve ourselves and develop our understanding”.
In this short video, Ernest describes his thoughts in more detail.
Our second keynote was Mario Klingemann whose artwork Interstitial Space was on display at Phoenix as part of the Art AI Festival 2019. He discusses how his work relates to classical art, such as paintings and portraiture, specifically discussing the work of Francis Bacon. In his talk, entitled Trapping the Accident, Mario describes how he believes we are entering a new stage in the use of AI for creative practice. The use of GANs (generative adversarial networks) are accessible to ever more artists although this leads to a dilemma: how is it possible to create an original language in this increasingly crowded creative space?
Image from Interstitital Space 2019
Mario describes AI as a new media form, much like photography, which he refers to as ‘neurography’, suggesting that creative questions cannot be addressed with any other media. In particular, he feels AI helps him as an artist by augmenting his own imagination, which he describes as limited as a consequence of being human. AI therefore enables him to find and create new ideas.
The BBC’s Amy Payne (reporter) filming with Britbot at the Art AI Festival 2019
In this short video, Libby describes her practice-based focus on AIs as chatbots, and how this creative form may potentially be used to disrupt narrowly defined narratives. For example, her artwork Britbot has evolved by learning from people that interact with it. In doing so, the concept of Britishness embodied within the original dataset, which she describes as having derived from a very middle class view of Britain, has dissolved and become more ‘blurry’.
Fabrizio Poltronieri, a researcher based in Leicester at the Institute of Creative Technologies, also discussed his creative practice. As one of the exhibitors at the 2018 Art AI Festival, in this video he also talks about his installation, the Love Apparatus. His practice with computers and AI is as a partner in creating art and solutions for society: he uses a mix of connectionist (neural networked) and symbolic AI in many different ways to achieve a particular solution. He sees them as symbiotic in their work with humans – he is passionate that the human remains ‘in the loop’ of creative practice and the finished artworks. Computers may well help us to create new possibilities in the world through the collaboration between the ‘creative minds’ of humans and machines. He stresses the importance of the festival in generating a space for a critical mass of creative AI practitioners to discuss its potentials for new artforms and practices.
Our final artist talk of the day was by Jake Elwes. As an artist using AI in his creative work, Jake talked not only about the creative successes he has had but also how he has used the failures and glitches generated by AI in his work. He shared some of the inspiration and outputs of his techniques for data capture, including “Dada Da Ta”, a piece in which he captured audio-visual content of the world’s most influential figures by cleverly embedding the story of their power and presenting it in a rather unique way. His presentation also highlighted a question that he explores through his practice: what is the role of the artist and how is their agency affected by AI? He emphasises that his aim is not to showcase the technology but to create something conceptually poetic that, although it can perhaps only be generated through AI, is not fundamentally about AI for its own sake.
Our day ended with a presentation by Luba Elliott on the ‘State of the AI Art’ – a summary of the breadth of creative practice used by artists including the work of our speakers and numerous others who are exploring its potential.