**2. Important considerations regarding AI in healthcare**

Digital bias is an important concept that is bound to become a mainstream consideration in the still very young AI era [41]. In this context, biases that are already present in various types and channels of "source data" have the potential to perpetuate existing healthcare disparities, resulting in a system that may be technologically more advanced but also one that continues to disenfranchise entire segments of the population [42]. Some applications of AI and machine learning in healthcare are starting to come to the forefront. Primary healthcare education and training, as well as the area of continuing education and training are also important areas where machine learning and AI can play a role. Currently, medical education involves a one-size-fits-all curriculum approach where everyone is given the same set of training/education (simulation-based, didactic, and otherwise) regardless of real-world clinical experiences and proficiency/competency levels. To this end, machine learning and AI in medicine can be used in different ways including personalizing training/education of healthcare professionals [43]. In this context, optimal training and education of healthcare professionals is a "big data" problem and via prediction of performance and knowledge/skill acquisition, maintenance and decay over time, it will be possible to personalize training for an individual provider [44–49].

With the advent of AI and machine learning in any field, there is always the worry that it will replace the jobs of professionals in the field. Although there has been tremendous growth and advancement of AI and machine learning in healthcare, bedside care providers are not at risk of replacement any time soon. In the near term, AI and machine learning in healthcare will primarily offer the potential to augment the performance of healthcare providers and simplify or support their clinical decision-making processes and clinical workflows. This will likely result in a reduction of workload by identifying patterns and trends in large electronic medical records databases and bringing to the forefront key information that will assist the provider in diagnostics and making the best treatment decisions for their patients. AI and machine learning will become extremely important in our fast-changing world and our continually evolving society, where staffing shortages of medical professionals are likely to remain a significant issue, with demographic trends working "against us" well into the future. Having AI and machine learning-based technologies which ultimately optimize the performance and efficiency of healthcare professionals is therefore urgently needed.
