“Who Wants to Write an Email?” is a Lir Academy show by Laura Allcorn and Jennifer Edmond. To call this Dublin Fringe Festival production, performed from September 9-12, just a “comedy show” or even a “play” would be misleading. It is really an interactive gameshow. For each round, Allcorn – who plays the quiz show host – invites one audience member to participate. Allcorn and Edmond have asked AI to generate emails using various prompts. The aim of the game is to fill in the blanks and guess what sentence the AI generated next in each email. There were three rounds in total, with multiple questions per round.
“My favourite prompt of the night when they asked AI to write Paul Mescal a flirtatious email.”
The concept is great. In a strange, almost meta way, the audience fulfils the role of AI, choosing the most likely answer for continuation of the sentence. The nuances of AI’s processes were written into the game and this made the show engaging. The prompts and emails were genuinely hilarious. Both Allcorn and Edmond’s senses of humour shined through, especially during my favourite prompt of the night when they asked AI to write Paul Mescal a flirtatious email. They sent the email through AI three times, asking it to make it more flirtatious each time: you can imagine how well that went. One of the best parts of the show was Jennifer Edmond’s expertise. Edmond is an internationally recognised researcher and Trinity associate professor of digital humanities. Drawing on these skills, the audience could call on Edmond for help in answering the questions. She was incredibly informative, and the parts where she was given the space to talk about AI gave the audience a break from the gameshow’s structure.
However, there were multiple problems with the show. While the style was fun, the concept did fall slightly flat after three rounds. The repetitive nature of the show started to show as the game continued. While the AI email prompt game was great, there needed to be something else or some other kind of game to introduce novelty.
Allcorn and Edmond attempted to inject the show with some structure. After each round of guessing, Edmondt would ask the audience if they were happy with the AI-generated predictions. Of course, the answer was usually no. Edmond would then give a “score” of how AI was deviating from audience expectations. When the score reached ten, the show ended. This part of the show seemed like it was supposed to go somewhere, but I had a sense that the gamemakers had not thought the concept through. I still do not know exactly what the point of this element was and how it was supposed to fit into the show: the execution required more consideration and planning.
While Allcorn was a great game show host, it felt like she was continuously holding back adding humour into the show. While some of her hilarious quips came through, she remained mainly silent except to coax the audience member playing the game. At times, this focus left the audience member hanging slightly. After all, whilst it isn’t the audiences’ job to be funny or entertaining, it was hers. While it is clear they want to let the humour of the prompts and the AI shine through, this was not enough to sustain the whole show.