Implications of embedded artificial intelligence - machine learning on safety of machinery
Sara Anastasi, Marianna Madonna, Luigi Monica
Abstract
The Artificial Intelligence (AI) and the Machine Learning (ML) is a rapidly evolving technology and up until recently has not been a subject of machinery safety. The purpose of this work is to evaluate how embedded artificial intelligence – machine learning can affect the safety of machinery and machinery systems in the development of their applications. This work can be useful to machinery designers to develop their particular applications as it describes how the new hazards, associated with embedded AI – ML, should be considered within the framework of the risk assessment process. The proposed study underlines the new dimension of complexity linked to artificial intelligence and machine learning that could lead to a revision of European legislation in terms of the introduction and/or modification of essential health and safety requirements (EHSR) in the Machinery Directive, in order to guarantee safety levels at least equivalent to those currently achieved.