World Health Organization: The Use of Artificial Intelligence in Medicine Can Be 'Dangerous' for Low-Income Countries

The World Health Organization has warned that the introduction of AI-based healthcare technologies could be 'dangerous' for people in low-income countries. This international health body states that it is essential for the development of these emerging technologies not to be solely shaped by large companies and wealthy countries. According to the WHO, if models are not trained on data from individuals in low-income areas, these individuals may not benefit adequately from algorithmic services. The WHO issued its first guidelines on AI in healthcare in 2021, but less than three years later, it had to update them due to the increased power and availability of 'large multimodal models' or 'LLMs'. These models, also referred to as generative AI, process and generate text and images, including the model that powers the popular chatbot 'ChatGPT'. LLMs can take one or more types of inputs such as text, video, and images and produce diverse outputs. In the medical field, these models are capable of generating clinical notes, filling out forms, and assisting doctors in diagnosing and treating patients. The WHO states that AI can play a significant role in improving the quality of healthcare, but only if those who develop, regulate, and use these technologies identify and fully address the associated risks. This international body warns that the exploitation of these tools should not be left solely to technology companies; rather, governments of all countries must make the necessary efforts to properly regulate the development and use of AI technologies, and civil society groups and individuals receiving healthcare should also participate in all stages of development and use, including oversight. In Iran, the use of AI in medicine has also recently come into focus. The Iran University of Medical Sciences claimed that an AI-based breast cancer detection system was developed at the university with '94 percent accuracy'. However, the necessary infrastructure for utilizing AI in medical science and what policies and programs have been developed for oversight and compliance with international standards remain unclear. Dr. Shahram Kordasti, a professor in the oncology department at King's College London and a hematology-oncology specialist, stated in an interview regarding the use of AI in oncology to Voice of America: 'Iran needs standardization and a clear method for diagnosis and referral processes to make the best use of existing techniques.' He mentioned that it is often assumed that because a topic is frequently discussed, it should be addressed quickly, but prerequisites must be established before implementation. The WHO's Secretary-General also emphasized that AI comes with serious challenges, including unethical data collection, cybersecurity threats, and the reinforcement of biases or the spread of misinformation. He urged countries to effectively harness its potential while simultaneously minimizing its risks through careful oversight. This report utilized sources such as 'Nature', WHO guidelines, and Iranian media.