Educating Without Extracting
An Anti-Colonial Approach to AI
By Femma Ashraf: Clinical Director and Co-Founder of the Pattern Institute
“The past few months have been personally challenging as I’ve found myself navigating racism in contexts where I would have preferred not to, professionally, socially, and structurally. These experiences haven’t just affected me as an individual; they reflect broader societal dynamics that are exhausting to confront and even more exhausting to explain.”
Femma AshrafClinical Director
So, I configured a GPT with the intention of curating resources and material created by Black and Brown thinkers, writers, and therapists. These voices are too often excluded from mainstream discourse, or worse, their work is scraped, extracted, and fed into algorithms without acknowledgment, consent, or compensation. The GPT draws from both societal and therapeutic lenses, centring the depth and breadth of perspectives that are frequently overlooked or silenced.
Initially, my focus was on creating a tool to educate white people, to offer an accessible, accountable space for learning. But attending a recent parliamentary roundtable on AI and inclusivity shifted my perspective. During that discussion, we explored how to meaningfully amplify Black and Brown voices in this rapidly evolving AI landscape. That moment helped me realise: this tool isn’t just about educating others. It’s about reclaiming space. It’s about resisting the erasure that so many of us have felt, even in fields where we hold deep expertise.
“At Patterns Institute, our ethos is grounded in education. This project is an extension of that work. I’ve renamed the GPT Beyond the Colonial Mindset to reflect the deeper intention, to unsettle inherited power structures, challenge the default narratives, and make space for voices and wisdom that predate and transcend colonial logic. You’re welcome to explore it here and share your thoughts.”
I also want to name the tension I hold in using a tool like ChatGPT. It’s resource-intensive, contributing to climate change through CO₂ emissions, water usage, and extractive infrastructures that disproportionately harm marginalised communities. We know, too, that many of the human workers who support these systems, such as content moderators in Kenya, are traumatised and underpaid, with little institutional support or recognition.
And yet, the alternative isn’t neutral either. The voices of Black and Brown people continue to be mined, appropriated, and left out of decision-making spaces. If we’re going to be in this space, and we already are, then I believe we must be in it differently. With intention. With accountability. With complexity.
This is just one attempt to do that.