There have been many cases of misdiagnosis in Kenya. The Kenya Medical Practitioners and Dentists Board blames machines for these cases. However, this is misleading since machines do not interpret the results. This is the role of medical personnel. Ironically, machines powered by Artificial Intelligence (AI) are learning fast and becoming so clever that it will be easier for medics to make more accurate decisions. Last week, I had the privilege of attending a Google AI Workshop, Making AI, in the Netherlands where experts discussed how to leverage AI to lower cases of misdiagnosis. The good news from the workshop is that advances in AI are narrowing the knowledge gap between what medics know and what they should know. Although AI is destined to disrupt virtually every sector, there is a more compelling case to leverage AI in medicine. Every medical facility has tons of records, which, if digitised, will help machines to make accurate assessment of disease patterns. Such patterns are perhaps what the medics need most for diagnosis, prognosis and prescription. It must be noted that machine ability to learn from millions of data sets is something no human can do. It brings years of experience into the hands of medics enabling them to focus on patient treatment. This, however, is necessary to assist medics especially in places where the ratio of doctors to patients is high, to make decisions that would otherwise take several days if not months. AI is, therefore, likely to help developing countries leapfrog several years into modern quality healthcare. This cannot be realised if we cannot, for example, facilitate AI research in order to tackle barriers to implementation. There is need to develop legal sandboxes to help test the application of AI. We have nothing to lose other than helping to stimulate the uptake of AI within priority sectors like healthcare. Diagnostic areas such as a medical laboratory would perhaps experience the greatest transformation of AI and improve the quality of care. Greater impact will be felt in countries that understand the disruptiveness of AI and work towards encouraging responsible data sharing to boost data availability for training AI systems. For instance, we cannot minimize Malaria/Typhoid misdiagnosis if we cannot digitise data from all of our health facilities and analyze it to build new knowledge on the treatment of these diseases. Even as we seek to leverage AI to solve many problems there are security concerns that that might fall into wrong hands, or that it might be used unethically. It is possible that human biases can be hardcoded into AI decisions, such as systemising inequality or infringement of individual liberties. There have been cases of blackmail from people who acquire others’ data. It is for such incidences that governments should promote constructive governance frameworks and build capacity within oversight agencies. Since innovation precedes regulation, governments should take the leadership role in showcasing how such emerging technologies can be used responsibly. Most of the health facilities in developing countries still fall within the public sector. That they can become an example of responsible AI use while at the same time giving knowledge to oversight agencies on what best practices they can use to regulate new innovations. In this era of rapid technological changes, there is need to develop a set of principles that can help guide use of transformative applications especially in saving lives while protecting citizens from those who use the same technologies for malicious purposes. Data protection is critically important at this stage when the entire world has realised that it is the oil of the future. Data is the main tool for building predictive models from diseases to weather patterns. Data is the future of our lives but as it is at the moment, there isn’t a clear policy guideline on how individual data can be used for a common good. AI presents a profound and promising prospect for improved healthcare provision. Rather than rubbish this emerging technology as a passing cloud, and going by the emerging research outcomes, we must give it a chance in order to understand it better through indepth research. The writer is an associate professor at the University of Nairobi’s School of Business. The writer is an associate professor at the University of Nairobi’s School of Business. The views expressed in this article are of the author.
