Classification of Data and Information in the AI Era: Towards Compliance and Efficient Data Management

In the era of increasing application of artificial intelligence (AI) and machine learning (ML) tools and services, AI agents, and agentic AI within organizations on data originating from multiple sources, it is becoming necessary for organizations to address compliance and efficient data management in an efficient and sustainable manner to remain competitive. The results of the case study are an evaluated and improved data management framework, applicable to many types of organizations; this framework includes AI/ML-guided classification of data and the addition of adequate cybersecurity and quality measures prior to using the data and information in AI/ML-based tools or services, AI agents, or agentic AI solutions. This paper highlights the need for a new or improved AI policy and suggests that organizations update or improve their strategy and processes to reap the benefits while avoiding exposure of confidential matters and non-compliance. The framework and its process, identify–protect/secure–prevent, can have significant practical potential and be useful for many organizations, such as universities, public entities, and companies, that possess large amounts of unclassified data. One of the organizations that was part of the evaluation was able to quickly classify more than 10 TB of data, and another was further able to delete/remove about 90% of its data, which was irrelevant and, to a large extent, unknown (i.e., dark data), enabling a transition to new infrastructure with significant cost savings, achieving compliance, and improving data management efficiency at the same time.