This month the WHO published a discussion paper on AI in health policy, and one idea in it deserves attention from anyone working in community care.
They call it "epistemic injustice": the tendency of AI systems to privilege quantifiable, data-rich evidence while marginalizing lived experience, local expertise, and community-based insight.
In our corner of the sector, that matters. The people who understand a client best are often the ones whose knowledge is hardest to fit into a structured field, the frontline worker, the family caregiver, the community that's been doing this work for years.
If a tool only "sees" what's easy to count, it can quietly steer decisions toward what's measurable and away from what's true.
The WHO's framing is a useful one: AI should augment, not automate. Humans stay responsible for framing the questions, judging the evidence, and weighing what the data can't capture.
Worth a read if you plan or fund care.
What this means for health organizations
When using AI for planning or reporting, pair what the data shows with input from staff, clients and community partners. Treat AI output as one input to a decision, not the decision.
