When President William Ruto launched the Social Health Authority (SHA) in 2024, the promise was historic: Universal Health Coverage (UHC) for every Kenyan, ensuring that harambees (community fundraising) for hospital bills would become a thing of the past.
But an investigative report by Africa Uncensored, Lighthouse Reports, and The Guardian reveals that the system’s digital backbone, an AI-driven means testing algorithm, is achieving the exact opposite, overcharging the nation’s poorest citizens while giving the wealthy a discount.
At the heart of the controversy is a method known as Proxy Means Testing (PMT). Because 80% of Kenya’s workforce is in the informal sector with no steady payslips, the government turned to an algorithm to predict household income.
Instead of asking what a person earns, the system asks about what they own: Does the house have an iron roof? Is there a flushing toilet? Does the household own a cow or a smartphone?
The investigation’s audit, however, found the system to be insensitive to real life. By reconstructing the algorithm, journalists discovered that the model frequently misclassifies poverty-stricken families as middle-income based on static assets. In one case, a family living below the poverty line was charged a premium twice as high as they could afford simply because they lived in a county labeled as prosperous and the parents had finished high school.
Perhaps the most damning revelation of the report is that the government was warned of these failures before the system ever went live.
A previously unpublished report from international consultancy ID Insight warned that the means testing was inequitable and highly likely to miscalculate contributions for low-income households. Despite this, the system was launched.
David Cawer, a health economist involved in the design, admitted that the system was intentionally weighted to favor inclusion errors. In simpler terms: the designers preferred to accidentally overcharge a poor person than to undercharge a rich one, fearing a loss of revenue. The logic was that a poor person could simply appeal the decision, but for many, that safety net has proven to be a mirage.
For Kenyans on the ground, these algorithmic glitches have life-or-death consequences.
- Florence, an unemployed mother in Huruma, was assigned a premium of Ksh. 6,600 per year, an impossible sum for her family. When she attempted to use the official appeals process, her request was summarily denied because her premium did not meet a minimum threshold for reconsideration.
- Community Health Promoters report that many are now being denied care at hospitals because the system lists them as ineligible or because they cannot pay the inflated premiums calculated by the AI.
As of early 2026, the SHA is facing a crisis of sustainability. While 30 million Kenyans are registered, only 5 million are actively paying their premiums. The lack of trust in the algorithm, combined with a bureaucratic transition from the old National Health Insurance Fund (NHIF), has left hospitals with billions in unpaid debt and patients stranded without coverage.
Former Deputy President Rigathi Gachagua recently claimed the system would collapse within months, a sentiment echoed by medical experts who argue the experiment has failed.
While government officials describe these as teething problems that will be sorted with time, critics argue that the very architecture of the AI must be scrapped.
“This is no longer just a question of design or data,” the investigation concludes. “It is a question of direction”. For the millions of Kenyans currently locked out of healthcare by a mathematical formula, the wait for a redesign is a luxury they cannot afford.
The full documentary is below: