Build-a-bot: take back your tools

A small number of powerful companies now control how most people interact with AI. Their tools are convenient, but not neutral - they are designed to keep you dependent, train on your data, make decisions about what intelligence looks like on your behalf, and consume extraordinary amounts of energy to run infrastructure that primarily serves their own growth.

This workshop is about engaging with this differently. We begin by looking at what a bot can actually be, from simple rule-based systems to generative text to locally-run AI, and how much wider the possibilities are than the chat window that has taken over the popular imagination.

You'll then design your own bot on paper: what it knows, how you interact with it, what it would never do, and who owns the data. From there, we bring those designs to life using Ollama, free open source software that runs large language models entirely on your own machine, configuring each bot with its own persona, purpose, and rules.

You'll leave with a bot running on your own machine, and the confidence to keep shaping it on your terms.

No prior coding experience required. Participants are encouraged to install Ollama before the session.

Workshop requirements: Laptop with Ollama installed.

About Mahalia Henry-Richards:

I am an artist, designer, musician, and educator. My background is in architecture and spatial design, and my practice now spans interface design, sound, installation, and live coding performance. I am a Senior Lecturer on the MSc Creative Computing at UAL's Creative Computing Institute, where I am also a doctoral researcher.

Across all of this, I am interested in how the tools and environments we use shape the way we think, make, and experience the world. Whether that's designing a space, performing with code, or teaching, I try to treat technology as something to question and reshape rather than accept as given.

My current research looks critically at generative AI and the power concentrated in the companies building it. Working with open source, locally-run tools has been one way of understanding these systems on my own terms, and of asking what, if anything, is worth building with them.