Fringe 2018: HumanMachine: Artificial Intelligence Improvisation


HumanMachine: Artificial Intelligence Improvisation is perhaps a first: the first Fringe show you could attend while simultaneously participating in a science experiment!

Improvisors Piotr Mirowski and Kory Mathewson are also researchers in the field of artificial intelligence. Mathewson is a familiar face on the Edmonton improv scene. After years of doing improv shows with regular ol’ humans, he tried an improv show with an AI in 2016. It failed so miserably, he spilled his guts about it in a blog post. Watching from London, UK, the Polish-born Mirowski saw the post and contacted Mathewson. Realizing their common improv/AI backgrounds and goals, they got together to create first a long-distance improv/AI show.

(A scheduling error had made their first long-distance simultaneous show a minor disaster, so perhaps it’s ironic that misreading their Fringe schedule has led me to post this entry during the last performance HumanMachine at this year’s Fringe.)

I caught one edition of HumanMachine. I’m not sure how much of it was part of their other performances, but for a show about what we can learn from failure, the performance I began appropriately enough when its small robot puppet, Raspberry, fell off its pedestal (about 8 minutes into a video of the performance on YouTube). Raspberry threw a couple of other curveballs at the technician, Riley, before Mathewson shouted from backstage: “Do you even want to do the show tonight, Raspberry?”

In explaining what we’ll see in the show, Mathewson and Mirowski break down all the elements of this robot improvisor: there is the physical mechanical being the programming that instructs it to move and “speak” (Raspberry doesn’t have a moving mouth; its “voice” came through an external speaker); and the content of its speeches. A human could provide Raspberry’s content, or an AI could pull from a database of lines sourced from, for example, terrible Hollywood movies.

They ran through a number of different permutations, including ones that involved guest improvisor Paul Blinov and an audience volunteer. Some trials were more successful than other. However, each little failure or unexpected success was an opportunity to reflect on all the things that go into a successful improv show, or successful human communications overall.

At the end of the show, audiences were invited to send feedback as part of Mathewson and Mirowski’s research project. Here’s hoping they replicate the experiment soon.

Fringe 2018 page:

Artificial Intelligence Improv website:


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.