An ethical messaging tool for movement organisations
What the grant built — and what it taught us
Loading report… 100% complete
Private by design • Built on movement values • Made by the communities it serves
GAIA
A fatal exception has occurred in the way progressive organisations use AI.
Grassroots groups are already using tools like ChatGPT — but every prompt sends the following directly into Big Tech infrastructure:
Corporate filters also hit "moral walls", flagging ordinary progressive language as "sensitive content" — quietly censoring the very groups we exist to support.
Press any key to continue building the alternative
PAQN is a small collective of queer people of colour — most of us raised in social housing, none of us the usual tech suspects. We came to AI from organising: years inside non-profit comms, writing the fundraising emails, building the donation pages, running the supporter lists — and we understand the tech underneath.
Abdul Boudiaf
Founder, PAQN • grant holder
Works at the join between community organising and technology — getting those two worlds talking.
Allaa Aldaraji
Strategic communications
A decade leading digital campaigns across the UK, Europe and the Middle East — storytelling, supporter engagement, fundraising.
Yasmine B
Creative technologist & researcher
AI, epistemology and data justice. Ada Lovelace Institute fellow; 100 Brilliant Women in AI Ethics 2022; teaches at UAL.
Batool Eldasouky Abdalla
Artist-researcher • UAL CCI
Associate Lecturer working across robotics, computational inequalities and experimental publishing.
What we set out to do
We did not want to accept that movements have to use AI on Big Tech’s terms. The grant was our chance to build an alternative — and to move from theory to a working tool.
Can we offer a route that is less extractive, more privacy-preserving, and actually aligned with movement values?
Phase 1 — Ethics Interviews
Six specialists, brought in to pressure-test the ethics and find the adoption barriers before we built anything.
“The most trustworthy version of this is a closed system. Local, offline, completely under the user’s control. That’s the only version where you can genuinely say: your data stays yours.”
“You can’t have an honest conversation about ethical AI without talking about energy and water. Large data centres use massive resources.”
“It’s easier to build momentum by making something tangible — a tool people can try — rather than by only telling people what not to use.”
Phase 2 — We Built It Offline, with Ollama
We tried cloud prototypes first. Fast and slick — and they failed the ethics test. So we built the local version.
The Problem That Changed Our Direction
Local kept data private — but a laptop-sized tool couldn’t keep pace with cloud models improving month on month.
Phase 3 — Gaia Messaging Tool Pilot
We piloted the messaging tool with two organisations already doing high-stakes campaign and fundraising work.
354K
▲ views
No Harbour, scripted with Gaia
~8M
▲ views
content via these workflows
+118%
▲ year on year
€23.6K → €51.4K
2
▲ real campaigns
high-stakes comms
The coalition organising against arms shipments through European ports drafted their launch video script with our support — it reached 354,000 views. Since July 2025, content made through these workflows has reached ~8 million views.
We supported their end-of-year fundraising. Year-on-year giving rose 118%, from €23,601 to €51,440 — with more content, a petition, and thousands of new subscribers.
ELSC — End-of-Year Giving
Year-on-year giving (€)
| Metric | Result |
|---|---|
| Content pieces published | 8 |
| Views across content | 411,568 |
| Petition signatures | 5,767 |
| New email subscribers | 4,927 |
Why it worked
It increased their capacity. A small team drafted faster, put out far more content than ever, and had room to test new framings and donor journeys — the freedom to experiment is what made the campaign succeed.
In Their Words
What the pilot organisations told us about working with the tool.
“We’d throw in our messy rough notes and campaign goals, and it would transform them into scripts and posts. We worried it would churn out generic, classic-AI content — but because we fed it a language & style guide, it kept our tone of voice consistent. In a coalition that really matters: it makes sign-off a lot easier. Saved us a bunch of time.”
“The real challenge is turning ideas into clear messaging that lands, repeatedly, without burning out. It increased our capacity. We’re a small team, so the year-end push usually leaves us running on empty — this time we drafted faster and had room to experiment. And because of how we handle the data, it felt safer than corporate AI.”
What Grew Out of It
The grant funded a messaging tool. The method turned out to reach much further.
A harder problem: confidential legal casework
Working with a movement legal team, we built an early tool that lets a lawyer search and summarise their own case files in plain language — with confidential client material never leaving the organisation.
It shows the method reaches well beyond messaging — to any group holding information it cannot hand to Big Tech.
The Wider Body of Work
The grant seeded a wider practice — built and delivered for movement organisations across Europe, with their consent.
No account, no lock-in.
No Harbour’s coalition includes Progressive International, Energy Embargo for Palestine and BDS — the work travels through the movement.
What It Taught Us
The grant settled an argument we’d been having with ourselves.
Ethics meant one correct technical choice — pick local, defend it forever.
Ethics is how you use these tools and how you handle the data people trust you with — and that judgement has to keep moving as the technology does.
“We come from the communities these campaigns are fighting for — so the difference between a tool that protects people and one that exposes them is not a design question. It’s personal.”
What We Want to Do Next
We’ve done a lot with a little. The next step is to make sure the value doesn’t stop with the groups who happen to know us.
Why now
Most advice about AI in the movement is still guesswork, and so many groups already use AI with no agreement on what’s safe to put into it — which is exactly where the harm starts. We want to keep building free tools the movement owns rather than borrows.
~97%
of young people in Britain already use AI tools — as trust in data protection falls
None of this would exist without the Ideas & Pioneers Fund — backing the project when it was still just an idea and a hunch.
Thank you, Aakash — for the questions that pushed this somewhere better, and for the support, guidance and passion.
Project
Gaia
By
PAQN
Status
DELIVERED
Progressive AI & Quantum Network • End of Grant Report • June 2026