Cover photo

Bringing refugee art onchain with Titles AI.

How a phone call 10 years ago lead to a custom AI model, trained on 200 pairs of recycled jeans.

deebee

deebee

An afternoon’s work? Ha!

Its the summer 2015, I get a call.

Do I want an afternoon’s work filming an art installation? Something about The UN, climate change and a dress made from a tent, that had once been the home for a family of Syrian refugees in Jordan.

Sure!

post image

Ten years later, I’m still working on the same project, at least it’s latest iteration - two custom-trained AI models by Titles.xyz created to bring refugee art onchain. On base.

Exactly how an afternoon’s work lead me here—via the streets of London, The UN, Glastonbury Festival, Jordanian refugee camps and The Venice Biennale—is another story for another time.

post image

This chapter begins in Dzaleka, a refugee Camp in Malawi.

Dzaleka

post image
post image

Dzaleka is home to over 50,000 refugees fleeing violence in neighbouring countries. It’s over-crowded with insufficient health services, water, shelter, and sanitation.

post image

Dzaleka Arts Lab evolved from a project by a team from University Arts London and thirty residents in Dzaleka.

A collaborative arts group empowering artists through traditional craftsmanship and storytelling, predominantly working with textiles.

post image

Initially my role in the project was to work remotely with a young refugee filmmaker in Dzaleka, Johnson Burume, to help him capture the emerging story and to produce a video documenting the fashion and music scene in the camp.

post image

In the end, many other things emerged and I got sucked into the project again.

The Map: (Mahali - 'place')

The first project to be co-created by the community was a collective narrative expressed through embroidery; each piece representing a personal story or logos of the creator.

post image

At the suggestion of the members of DAL the pieces were brought back to the London College of Fashion and combined into a map of Africa. A collective co-created artefact grounding the narratives of over 30 people into a representation of place.

post image

When I saw the piece for the first time I realised it deserved a life beyonds it’s physical form.

Energy DAO

I was interested in finding a way to bring bring the work of Dzalaka Arts Lab onchain, to help generate income for members through NFT sales.

I’m a member of Energy DAO, a creative grants collective built on Nouns Builder, that funds arts and cultural projects on the Superchain. I submitted a proposal for 1 ETH to prototype ways to bring Dzaleka Arts Lab onchain, and got the nod.

The DAL community decided on a theme for their next collective artwork—Home.

It wasn't what any of us expected.

Pockets: ( NYUMBANI - 'home')

post image

 ‘Pockets’ is constructed from the pockets of over 200 pairs of recycled jeans originating from a local scam. Originally planning to make tote bags from second-hand denim skirts, the group received shorts instead. Forced to adapt, they creatively reused the pockets to build an epic tapestry, over several months of work.

The pockets symbolise the reality that, when fleeing home, as many of them had, everything meaningful you carry fits in your pockets.

An amazing piece of work, but how do we bring it onchain?

Co-co-co-creation with Titles.xyz

Both artworks are collaborative physical pieces that contained the narratives of the people who had made them. We wanted to carry this theme of co-creation onchain.

So we turned to Titles to help create two artist-owned AI models that would allow anyone to create their own AI -generated work inspired by the original physical art works, and to collect the work of the collection to become part of the evolving story.

The Process

The first step was to create, mint and share the training data for the models.

post image

These were essentially curated and cropped photographs of the details of the artworks that we wanted the model to use. The team at Titles then trained the models, tested and tweaked them to get the best results from a range of different prompts.

The Models

You can mint you own creations with the two custom trained models, DAL receives a portion of the fees from each mint to help support the community.

NYUMBANI - 'home'

The DAL community have created a collection of 10 pieces using the Dzaleka Stitch: The Map model.

post image

post image

Each piece is a collaboration of three members, with the prompts seeded from a three word slogan and corresponding image inspired by the slogan.

All mint funds received will go towards supporting the community and their ongoing work.

Become a supporter of DAL

post image

The images from the training data are available as 1/1 NFTs to collect for 0.05 ETH on base HERE. All proceeds go to support the DAL community. All proceeds go towards funding and supporting the community directly.

DAL Website

deebeeFarcaster
deebee
Commented 2 months ago

Find out how we brought refugee's art onchain, on @base with @titles and an AI model trained on a tapestry made of 200 pairs of recycled jeans. https://paragraph.xyz/@deebee/bringing-refugee-art-onchain-with-titles-ai

Bringing refugee art onchain with Titles AI.