Machine learning AI systems and the virtue of inventiveness

Abstract

While the subject of artificial intelligence as an inventor has centered on statutory compliance, a more fundamental question is whether AI can technologically fulfill the role of an inventor. In other words, does the current state of the art merit the virtue attribution of being an inventor. To address this question, an analysis of the technology domain is given. The emphasis of technological analysis is centered on what constitutes the effort to produce an invention. The analysis then follows with the underlying assumptions of AI with specific focus given to machine learning. This basic groundwork provides enough information on which to analyze the challenges faced by machine learning systems on the field of inventing. The challenges form a set of requirements and evaluation criteria for design and disclosure of AI machine learning systems. Even under the assumption of overcoming the challenges presented is the notion of autonomy in inventorship claiming. The analysis leads to the possibility of determining execution autonomy as levels of machine learning processing. The challenge presented here, along with the autonomy criteria, poses a new evaluation criteria that can help answer the question on whether to grant inventorship status.