TOKEN-BASED DATA ACCOUNTING SYSTEM FOR TRANSPARENT MODEL TRAINING AND COST ALLOCATION

Abstract

The present report examines the idea of a Token-Based Data Accounting System, which is to be implemented to increase the transparency of the process of training the AI/ML model and providing the opportunity to properly allocate the costs. The more organizations implement artificial intelligence (AI) and machine learning (ML) technologies, the more complicated it consists of managing data utilized to train the model, and also the costs associated with it. The traditional systems can be characterized by a lack of transparency in the manner in which the data is consumed, in addition to their way of distributing the costs among different stakeholders. A tokenized data accounting system uses tokenization, blockchain technology, and smart contracts to develop a non-transparent, efficient, and fair system of handling such resources. The system will monitor data contribution, apportion costs of data usage, and they will be compliant with ethical and regulatory requirements.