Etisk problematik vid implementering av maskininlärning i sjukvården

Sammanfattning: Technological advances enable new applications in a number of industries. Artificial Intelligence, more specifically Machine Learning, is technology that is expected to have enormous potential in healthcare. Healthcare is constantly facing critical challenges but Machine Learning is anticipated to make it safer and more efficient. However, the implementation of Machine Learning also presents healthcare with ethical dilemmas. Based on the European Commission's Ethics Guidelines for Trustworthy Artificial Intelligence, this study explores those dilemmas in the context of Swedish healthcare. An in-depth qualitative case study is conducted in which doctors are interviewed and data is collected. The collected data is analyzed and with the help of the European Parliamentary Research Service's compilation of ethical problems within Artificial Intelligence, the study identifies six ethical problems from a doctor's perspective. These ethical problems the study identifies may arise when implementing Machine Learning in healthcare are “Bias,” “Transparency,” “Relationships,” “Inequality,” “Privacy, Human Rights and Dignity,” and “Why trust is important.”