When Do Data-Driven Systems Exhibit the Capability to Infer?
A new research paper proposes a framework to define and measure the inferential capabilities of data-driven systems, particularly in the context of the European AI Act. The paper aims to clarify the regulatory status of systems like credit scoring models, which may or may not be considered AI under the Act depending on their inferential capacity. By analyzing entire data processing workflows and the role of human experts, the research seeks to provide regulatory clarity and identify areas needing further definition. AI
IMPACT Provides a framework for understanding AI system capabilities under new regulations, potentially impacting compliance for data-driven services.