MozFest 2021: Learning to be thankful with the Gratitude Machine

Mark Kirby, Community Librarian (West) for Leeds Libraries, explains how The Gratitude Machine session at MozFest 2021 provided him with an entry into machine learning.

Part of my pre-MozFest plan was to become better acquainted with concepts of AI and machine learning through the festival. After attending a couple of sessions, my understanding was still pretty hazy and lost in abstraction, until one session in particular helped lift that haze. That session was The Gratitude Machine.

“How can you take machine learning and AI to assert joy?”

The session was led by Jo Kroese, designer of the Gratitude Machine, and the question above is one that drives Jo’s work. The Gratitude Machine is a machine learning interface based around the decidedly human and anti-capitalist ideas of gratitude and joy. The premise is simple: you type to enter an expression of gratitude beginning with the words ‘Thank you for…’ into the Gratitude Machine, and the machine responds with a series of thankyous in return. As Jo says, “The Gratitude Machine is learning how to be thankful”.

Me: Thank you for nettles and the delicious spark of their sting.

The Gratitude Machine: Thank you for fires still dancing in the wind and sparkles in the rosebush.

It helped that, the previous day, I’d attended a metta-bhavana meditation session. In this meditation, thoughts of loving kindness are sent out to others. Metta-bhavana meditation finds an echo in the process of The Gratitude Machine, and being able to bring that personal experience into the session gave it additional resonance.

“AI becomes what we feed it.”

The Gratitude Machine enacts this idea beautifully. It places responsibility on us to feed it thankyous of sincerity. In doing so, it also leaves the machine vulnerable to being corrupted. This sense of responsibility, of vulnerability, of the need to be aware of the processes by which AI is constructed, provided me the gateway into machine learning and AI that I was seeking from MozFest, and for that, I’m very grateful.

Thank you for reading this messy set of thoughts. Thank you for giving the Gratitude Machine a try:

https://gratitudemachine.io