A.L.Ex and the ImpRobots: an AI Show for Kids

Interview By Lou with Piotr Mirowski from Deep Mind

What an honour it is to interview Piotr Mirowski, a scientist who works with A.I. for Deep Mind, a scientific company that is becoming increasingly known. I had not expected that! Here, we talk about the show and also some of the more burning questions of the day about AI, such as when it comes to jobs etc. I feel the answers are fascinating and important.

Let’s welcome to Bookmarks and Stages Piotr Mirowski

1. What can people expect from Artificial Intelligence Improvisation and from A.L.Ex and the ImpRobots: an AI Show for Kids?

Artificial Intelligence Improvisation and A.L.Ex and the ImpRobots! are two interactive live experiences featuring professional actors, cute real robots (an Aldebaran Nao), and various flavours of artificial intelligence on stage.

Artificial Intelligence Improvisation is our pioneering improv comedy show with AI: it is for a general audience and addresses conversations about human agency (some actors wear augmented reality glasses and are controlled by AI that sends them lines) or about deep fakes, and at the same it time showcases the ingenuity of human improvisers.

A.L.Ex and the Improbots! is a show for kids and the whole family where fellow young audience members learn fun facts about science and are invited on stage to co-create stories with robots.

In both our shows, Artificial Intelligence Improvisation, and A.L.Ex and the ImpRobots, language models give strange suggestions to human improvisers, giving them an opportunity to react, to incorporate the strange material and then to shine and to demonstrate their talents and sense of humour.

2. Why did you decide to cross science with comedy with an AI robot and what challenges did you have in setting this up?

Our primary aim is to leave audiences, and the younger generations, feeling empowered to critically engage with AI, and to directly explore for themselves through co-creation with the tool, rather than passively.

The show’s creators, Piotr Mirowski (that’s me!), Boyd Branch and Kory Mathewson are academics who are passionately engaged in communicating about science. We are also theatre actors who want to bridge disciplines. Their challenge is to explain, succinctly, the complexity of a fairly complex topic to the audiences, to give them back control over a key technology in their lives. When we started the shows, back in 2016, everything sounded new, from machine learning, to patterns, to biases in data. Today, most people have already tried experimenting with text and image generation via various apps.

The real challenges now, in a way, are to manage the very high expectations about what AI can do! In our very fast paced show, AI tries to react, live and in a fraction of a second, to the mayhem on the stage, and speech recognition and large language models really struggle to keep up!

3. How does it feel bringing a groundbreaking AI Improv show to the Edinburgh Fringe Festival?

Kory and I actually brought Artificial intelligence Improvisation to Edinburgh in 2017. Kory (at the time studying for a PhD in robotics at the University of Alberta) and I were doing a duet with a twelve-inch robot and performed for a week at Surgeons’ Hall. The show was very experimental and very rough around the edges, but our friend Colin Mochrie (from Whose Line Is It Anyway?) came to see our show and had some good words about us.

We came back to Edinburgh last year and this year, to perform in a larger space at the Gilded Balloon. We encountered extremely supportive production and tech that made our tech-heavy show a (relative) breeze to get in and get out each day.

As a personification of the AI, our new robot is a bit larger this year, but it is extremely cute and gets lots of enthusiastic reactions from audiences or even from crowds when we take it out to flyer with us.

We’ve also seen amazing productions in Edinburgh that involved robots and AI. Shows that really marked us were the production of Spillikin at the Pleasance in 2015, of Siri by Laurence Dauphinais at Summerhall in 2017, or Robo Bingo by Foxdog Studios last year at Underbelly.

4. How much of an influence do you think AI will have on humans in the future?

I see AI primarily as a tool for search and discovery. We have seen inspiring examples of AI tools that can make predictions about the structure of proteins, predictions which can then be verified experimentally by chemists and biologists. I have worked on using AI for making weather predictions, helping expert meteorologists refine weather forecasts, with weather agencies now evaluating how AI can help predict the trajectory of hurricanes. We know artists who have been experimenting with AI tools and exploring the glitch aesthetic of their input to integrate this strange instrument in their process and create unique art.

5. Since working in the field of epileptic seizure predictions, mapping on smartphones and more for the likes of Bing etc, what made you decide to now join Deep Mind to work with AI in the artistic field in co-creations on stages?

The work of co-creation with AI for live theatre performance is done in my spare time and through my theatre troupe, Improbotics.

In my current job, I have worked on navigation, and in recent years, on weather forecasting and applications to climate modelling. However, my exposure to the theatre world inspired me to also focus on the ethical concerns when using AI in the context of the arts, and to evaluate the usefulness of language models as a tool for creative writing. Two years ago (before large language models became so popular), I ran workshops with screenwriters and playwrights trying to write with AI tools: their contrasting reviews were published at a conference on human-computer interaction. At the previous Fringe Festival, I took advantage of my presence in Edinburgh to interview comedians who had been using AI, to understand if large language models are aligned with the cultural values of comedians, and published findings at an AI ethics conference (spoiler alert: the comedians were not impressed).

6. Perhaps the question everyone really wants to know, since everyone talks about it, is: Many people in most industries are worried about their jobs as AI advances ever forwards, how does that make you feel and do you ever worry about your own job in this context?

I believe in the need for empathy and dialogue between developers and the rest of the civil society, in order for us to realise AI’s potential as a tool that benefits everyone.

My personal belief is that we all tend to underestimate the complexity of human activities (in particular when we are talking about other people’s jobs…) and that we forget the need for human connection and for sharing our lived experiences – which simply cannot be automated. The latter point was apparent when we interviewed comedians who had tried using AI for comedy writing.

For these reasons, I do not see AI as a substitute for work or for the process of writing, thinking and creation. I know that there is more to writing than merely putting words on a page, and there is more to computer science and engineering than merely writing lines of code. I believe there are better uses of AI than for the “automation of mediocrity” and am confident we can build a future where AI tools are used to help, not replace human activity.

Tickets here via The Edinburgh Fringe website: https://tickets.edfringe.com/whats-on/a-l-ex-and-the-improbots-present-an-ai-show-for-kids