Governance, Accountability and Post-Deployment Monitoring Preferences for AI Integration in West African Clinical Practice: A Mixed-Methods Study
This mixed-methods study of West African clinicians and technical experts reveals a strong preference for independent regulatory oversight, transparent accountability mechanisms, and real-time monitoring to ensure safe and equitable AI integration in clinical practice, while highlighting significant concerns regarding vendor control and clinician liability.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine the world of medicine as a massive, bustling kitchen where chefs (doctors) are trying to feed a hungry city. For years, they've cooked with their own hands, relying on experience and intuition. But now, a new kind of sous-chef has arrived: Artificial Intelligence, or AI. Think of AI not as a robot with a mind of its own, but as a super-fast, super-smart recipe book that can scan millions of meals in a second to tell you if a dish is going to taste great or if it's going to burn. In rich countries, these digital sous-chefs are already helping chefs spot tiny mistakes in recipes (like spotting a disease early) that humans might miss. But in West Africa, where the kitchen is often crowded, the ingredients are different, and the chefs are stretched thin, simply dropping a fancy new recipe book on the counter isn't enough. You need a system to make sure the book is actually right for your kitchen, that it doesn't give bad advice when the weather changes, and that if a meal gets ruined, everyone knows who is responsible. This is the story of how to build a safe, fair, and trustworthy kitchen for the future.
This paper is like a giant town hall meeting where the chefs of West Africa, along with the people who built the recipe books (the tech experts), sat down to ask a very important question: "How do we want to run this new kitchen?" The researchers didn't just guess; they asked 136 doctors and interviewed 72 experts to find out exactly what they need to feel safe using AI. They discovered that the chefs are terrified of one thing: being blamed for a mistake the recipe book made. They are also very suspicious of the companies that sell the recipe books, fearing those companies will try to control the kitchen too much. Instead, the doctors want a strict, independent referee—someone who isn't the chef, the hospital boss, or the recipe book seller—to watch over everything. They want to see the recipe book's performance on a live screen, like a sports scoreboard, so they can spot problems the second they happen, rather than waiting for a report at the end of the year.
The study found that when it comes to who should be the referee, the doctors overwhelmingly chose an Independent Body. About 40.4% of the doctors said this is the best option, giving it a trust rating of 4.3 out of 5. In contrast, they barely trust the companies that make the AI tools to watch themselves; only 3.7% of doctors wanted the "Vendor" to be in charge, and they gave these companies a low trust score of just 2.4 out of 5. The doctors also voted that 94.1% of them think having a clear path for who is responsible when things go wrong is absolutely essential. They are worried that without this, they will be unfairly blamed for AI errors, a fear shared by 76.5% of the respondents.
When it comes to how they want to check if the AI is working, the doctors don't want to wait for a slow, annual report card. They want a Real-time Dashboard, which 41.9% of them chose as their favorite way to monitor the tools. This is like having a live feed of the kitchen's temperature and the quality of every dish as it's being cooked, rather than waiting to taste the food weeks later. The experts interviewed agreed, saying that AI is like a car that can slowly lose its steering over time; if you don't keep checking it, it might drift off the road. They emphasized that the monitoring needs to be continuous, checking if the AI is still accurate, fair, and safe for everyone, including people in different towns or of different ages.
The paper also highlights what happens when the AI makes a mistake. The experts suggest a clear "stop and fix" plan: if the AI starts acting weird, you pause it immediately, figure out why (was it bad data? a new disease?), fix the model, and tell everyone what happened. The researchers suggest that for AI to work in West Africa, we need a multi-layered system: an independent referee, a live scoreboard for performance, and clear rules about who is to blame if something goes wrong. The study suggests that without these protections, doctors might be too scared to use these powerful tools, and the promise of AI helping patients could be lost. The authors are careful to note that while these findings are strong, they are based on the preferences of the people they asked, and the results might look slightly different in other parts of West Africa where the doctors weren't as represented in the survey. But the message is clear: for AI to be a helpful sous-chef, it needs a referee it can trust, and it needs to be watched closely, every single day.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.