Challenges and Solutions for Artificial Intelligence Adoption in Healthcare – A Literature Review

Even though Artificial Intelligence (AI) solutions are already used and have additional potential to revolutionize healthcare (HC), its application in practice tends to face limitations. Objectives of this research are to identify and summarize knowledge on the challenges and solutions for the adoption of AI in HC by a literature review of articles (N = 19) published between 2018 and 2022 in Scopus, Science Direct and BMC Health Services Research using the key terms “challenges”, “solutions”, “artificial intelligence”, “AI”, “adoption”, and “healthcare”. The study reveals several limitations of AI implementation: the lack of transparency in AI decision making, difficulty to trust and use the results; obtaining high-quality properly labeled medical data; concerns about data privacy, regulation and security and hesitancy of HC professionals and patients to trust and use AI-powered systems. To overcome these challenges, studies suggest to apply the solutions: justify the application of AI-powered systems in the context of digital transformation of HC - value-based HC and change management strategies with the aim to benefit patients and improve overall health outcomes; educate stakeholders and involve them in the implementation process to overcome black-box decision making; teaching AI fundamentals and applications to ensure HC professionals can effectively use and understand AI systems; address the challenge of data availability by federated learning and secondary use of health data enabling researchers to discover new knowledge and breakthroughs in medical science. Proper regulation is also essential to ensure the safe and ethical development, testing, and use of AI in HC.