We all tell ourselves stories about AI: how it works, who it is for, and what it can do. These stories hit without us noticing through films, books, memes, quietly shaping our assumptions long before we touch a tool. They arrive pre-narrated by filmmakers, newsrooms, by the companies selling it. AI will save us, AI will end us, AI is coming for your job.

In 2023 we came across the term "Slow AI" in this article about Timnit Gebru, and rather than nod along from afar, we wanted to work out what it could mean in practice. So we gathered 70-odd people on Miro to figure it out together: thinking out loud, dumping links, disagreeing, clustering. And 3 years on, that one session has become 3 reports, 4 zines, and 3 commissioned artworks, with 59 named authors between them.

But the news kept newsing: AI is bad for the power grid, data centres are being built on drought-hit lands, AI as a weapon of war. There's always something new to un/learn. Turns out, sensemaking isn't a project you ever finish. It's a forever practice - best done for yourself and with each other.

Thankfully, AI haters, hotties, and weirdos around Europe have joined us in this community practice. Ramla Anshur is doing it in London with and for the British Somali diaspora. Institute of Machine Unlearning is doing it in Lisbon. Studio Imposter and Plateia is doing it in Athens. DigitalFems is doing it in Barcelona (with a playbook in Spanish and Catalan coming soon!).

Insights from this work have taken root in new, artist- and community-led AI projects. Like Maria Than who has carried Slow AI into her Adobe Creative Resident in AI + Digital Worlds residency at London’s Victoria and Albert Museum. Or Chipo Mapondera who’s built on Ancestral AI to make Cry to the Water - a language model trained on oral histories and Tunisian climate data.

Every time someone makes something with AI, or runs their own session and shares what they found, we all get a firmer grip on what the heck this actually is. We help each other imagine what it could be used for - and, maybe just as usefully, what it probably can't.

Seen Slow AI out in the wild? Reply and tell us about an initiative or project you've come across - we're always collecting them.

We want to do more of this work: community sensemaking, hacking and subverting commercial AI tools so that they work for us, DIY AI. We’ve prepared a 2-year programme of activities called AI for the Rest of Us for a Stimuleringsfonds grant. Read more in our application here → If you know a person or an org who'd help fund this, hit reply — we'd love an introduction.

This newsletter was supported by Imagine Intel, the zine and cultural artefact published by Mozilla Foundation to accompany the Imaginative Intelligences Assemblies work led by the Foundation’s Creative Futures program. Check out the work of The Counterstructural Commons, Mozilla Foundation’s cultural R&D residency at Rhizome.