The purpose of this research and its corresponding thesis is to give the reader a deeper understanding of what the ethical risks with using machine learning are for a decision-making purpose in clinical examination and diagnostics in the medical field. The motive to why identification of these said ethical risks are essential, is also for the purpose of giving the reader an understanding of how these risks can be reduced. The information is gathered through earlier work that shows relevance but also through qualitative interviews with people working in the fields of machine learning. Through comparison of the results from the interviews with earlier work and other valid and legitimate sources, the result has been derived. The results concluded of ethical risks as a result from insufficient data, workers putting all of their trust in the result from machine learning algorithms, bias caused by humans when constructing the algorithms. Although there are many risks, there are also possibilities which reduce ethical risks consisting of improvement of technology and better implementation of machine learning in a working environment.