Health policies driven by AI are denying support to those in need in one of the poorest nations globally.

      A problematic AI system is escalating medical expenses for Kenya's most vulnerable families.

      Kenya had pledged to enhance access to more affordable healthcare. However, a recent investigation has uncovered that its algorithm-driven system is complicating life for the very individuals it aimed to assist. According to reports from The Guardian, Africa Uncensored, and Lighthouse Reports, the country's new Social Health Authority system employs a predictive machine-learning algorithm to determine how much individuals should pay for public health insurance.

      This system was initially launched in October 2024 as a part of President William Ruto’s commitment to broaden healthcare access for Kenya’s extensive informal workforce.

      How the algorithm is adversely affecting Kenyans

      A nurse cares for a young patient while a guardian observes in a hospital ward. Speak Media Uganda / Pexels

      The issue lies in the way the system assesses what individuals can afford. Kenya’s SHA system relies on proxy means testing, a technique that estimates income based on household characteristics such as roofing materials, toilets, livestock, family size, and other living conditions. The investigation revealed that the system has been inaccurately overestimating the incomes of low-income households while underestimating those of wealthy individuals.

      One SHA volunteer shared experiences from visiting homes in Nairobi, where residents already struggling to afford food were faced with premiums vastly exceeding their financial capabilities. Some reported charges equal to 10% to 20% of their minimal incomes, according to the findings.

      When a bill hinders treatment

      @AfUncensored @LHreports sought to confirm that the implementation of Kenya's new healthcare system would ensure a fair process for determining insurance premiums. Instead, we discovered an algorithm that overcharges those least able to pay while undercharging the affluent…— John-Allan Namu (@johnallannamu) May 4, 2026

      The implications are significant, and the outlook appears grim. Kenyans without private insurance who cannot afford their SHA premiums face the risk of being denied access to healthcare facilities or receiving exorbitant hospital bills. The report includes accounts of critically ill individuals missing out on treatment because the system indicated they owed more than they could afford. One single mother reported her monthly payment was set at 3,500 Kenyan shillings, while others noted significant increases from their previous payments under the older system. Consequently, the new policy is affecting lives.

      Although Ruto characterized the system as AI-driven, the report clarifies that it does not utilize generative AI like ChatGPT. Instead, it employs predictive machine learning based on a decades-old policy instrument that has faced long-standing criticism for misidentifying those deserving of assistance. Many have deemed this system flawed and inequitable, even prior to its implementation.

      Over 20 million individuals are registered for SHA, but only around 5 million consistently pay their premiums. Hospitals are also reporting considerable deficits due to unpaid reimbursements. This scenario underscores the risks associated with algorithm-based welfare systems.

      Vikhyaat Vivek is a tech journalist and reviewer with seven years of experience covering consumer hardware.

      Music is not the only sector suffering from low-quality output; AI has also significantly impacted podcasts.

      A.I. podcasts are turning audio platforms into content factories with minimal effort.

      AI-generated material has already overwhelmed video platforms, gaming discussions, software development, and search queries. Now, this same low-effort machine-generated content is infiltrating the podcasting arena.

      While music often takes center stage in discussions about AI's impact, the problem with podcasts may be less noticeable and more challenging to rectify. AI technology now enables the rapid creation, uploading, and monetization of entire shows, far outpacing traditional podcast studios.

      The renowned Oak Ridge lab has developed a portable device capable of detecting live GPS spoofing.

      Criminals can manipulate your GPS, making everything appear normal. Researchers have now created a solution.

      We place our trust in GPS systems much like we trust gravity; it functions reliably and guides us to our destinations. However, what if someone could deceive it into providing false information, leaving you completely unaware?

      Unlike jamming, which inundates your GPS with interference and alerts you to a problem, spoofing transmits deceptive signals that seem entirely authentic. You might be tracking your vehicle or a delivery, believing everything is in order, when, in fact, the shipment has been rerouted elsewhere, leaving you oblivious.

      I utilize AI daily, and here are three reasons I chose Claude over ChatGPT.

      I make use of AI every day, necessitating a tool I can rely on consistently rather than occasionally use. Eventually, I realized that for that kind of reliability, I’d need to invest in a subscription. The true challenge emerged when I had to make a choice. It boiled down to ChatGPT and Claude. Having used ChatGPT extensively, it already understood my thought process and needs, making it a comfortable option. However, as I explored Claude's capabilities, the decision became more complicated. The choice was no longer obvious.

      I deliberated