How to Shift AI Use in LMIC Healthcare?
AI is accelerating the idea of task shifting in global health. However, there is no shared way to say what was shifted, to whom, with what intelligence, under what fallback, and on what evidence.
WHO's task-shifting framework describe cadres and shift types, and wasn't built for AI tools. Two deployments both called "AI decision support for community health workers" can differ in material ways, frustrating donors who want to fund them, regulators who want to compare them, and Ministries who want to certify them.
There are five questions that we need to answer:
What tasks are shifted?
To whom are they shifted?
With what intelligence and supervision processes?
Under what fallback?
On what evidence?
SHIFT-AI is a framework built to force those answers. It will be submitted to The Lancet Digital Health. This session tries to break it first.
What to expect
Participants will engage with different deployment models. Practitioners will advocate their interpretations. Participants and practitioners will disagree, and we will all learn where AI can shift tasks (or not) in real deployments.
You'll leave with a framework you can apply now, the knowledge of where disagreements happen, and the preparation task-shifting discussions you'll have in the future.
Who this is for
Leading technology and development practitioners, global health funders and policymakers, and organizations operating in LMICs that are already using AI or trying to determine how to deploy it responsibly.
Who is leading the discussion
Soumyadipta Acharya, Johns Hopkins University
Subhashree Dutta, the^delta
Temina Madon, Agency Fund
Fred Tan, Hewlett Packard Enterprise
Jordan Shuff, Visilant
Moderated by Nikesh Shah, Intelehealth