Background. Artificial Intelligence (AI) and Machine Learning (ML) technologies make predictions from data to learn and adapt performance on a given …

The U.S. Food and Drug Administration (FDA), Health Canada, and the United Kingdom's Medicines and Healthcare products Regulatory Agency (MHRA) have jointly identified 10 guiding principles that can inform the development of Good Machine Learning Practice (GMLP). These guiding principles will help promote safe, effective, and high-quality medical devices that use artificial intelligence and machine learning (AI/ML). Artificial intelligence and machine learning technologies have the potential to transform health care by deriving new and important insights from the vast amount of data generated during the delivery of health care every day. They use software algorithms to learn from real-world use and in some situations may use this information to improve the product's performance. But they also present unique considerations due to their complexity and the iterative and data-driven nature of their development.

Since 2018, Health Canada has undertaken an initiative to adapt its regulatory approach to better support digital health technologies, specifically medical devices. Key focus areas include artificial intelligence, software as a medical device, cybersecurity, medical device interoperability, wireless medical devices, mobile medical apps and telemedicine. To meet this goal, Health Canada established the Digital Health Division under the Medical Devices Bureau and has been increasing its efforts to build in-house expertise. On October 27, 2021, Health Canada, the US Food and Drug Administration (FDA), and the United Kingdom's Medicines and Healthcare Products Regulatory Agency (MHRA) jointly published the Good Machine Learning Practice for Medical Device Development: Guiding Principles. The document consists of 10 guiding principles to help promote safe, effective, and high-quality use of artificial intelligence and machine learning (AI/ML) in medical devices.

We define a Causal Decision Problem as a Decision Problem where the available actions, the family of uncertain events and the set of outcomes are related through the variables of a Causal Graphical Model $\mathcal{G}$. A solution criteria based on Pearl's Do-Calculus and the Expected Utility criteria for rational preferences is proposed. The implementation of this criteria leads to an on-line decision making procedure that has been shown to have similar performance to classic Reinforcement Learning algorithms while allowing for a causal model of an environment to be learned. Thus, we aim to provide the theoretical guarantees of the usefulness and optimality of a decision making procedure based on causal information.

SAP has released its guiding principles for artificial intelligence (AI). Recognizing the significant impact of AI on people, our customers, and wider society, SAP designed these guiding principles to steer the development and deployment of our AI software to help the world run better and improve people's lives. For us, these guidelines are a commitment to move beyond what is legally required and to begin a deep and continuous engagement with the wider ethical and socioeconomic challenges of AI. We look forward to expanding our conversations with customers, partners, employees, legislative bodies, and civil society; and to making our guiding principles an evolving reflection on these discussions and the ever-changing technological landscape. We recognize that, like with any technology, there is scope for AI to be used in ways that are not aligned with these guiding principles and the operational guidelines we are developing.

Artificial intelligence (AI) technologies--and the data driven business models underpinning them--are disrupting how we live, interact, work, do business, and govern. The economic, social and environmental benefits could be significant, for example in the realms of medical research, urban design, fair employment practices, political participation, public service delivery. But evidence is mounting about the potential negative consequences for society and individuals. These include the erosion of privacy, online hate speech, and the distortion of political engagement. They also include amplifying socially embedded discrimination where algorithms based on bias training data are used in criminal sentencing or job advertising and recruitment.