Stanley Greenstein (Stockholm U Law) has posted “The Data-Driven Approach and the Legislative Process” on SSRN. Here is the abstract:
The age of digitalisation and hyperconnectivity has resulted in big data. This has promoted the development of complex machine learning technologies for extracting novel insights from data for decision-making purposes. These emerging technologies, incorporating elements of artificial intelligence, have enabled a new approach for utilising data for decision-making purposes, this trend referred to as the ‘data-driven’ approach. It has elevated the role of data from decision-supporting to de facto decision-maker, humans increasingly relinquishing this role to the self-learning models constructed from this data. The data-driven approach promises great potential for society: effectivity, objectivity, profitability, transparency, decreased bias and novel insights into human behaviour. There are calls to utilise the data-driven approach within the legislative and regulatory process. However, for all the benefits, the data-driven approach is not without dangers – processing data is complex, where bias, context dependency and a blind faith in the data can lead to flawed outputs jeopardising society’s well-being. This article examines the technical basis of the data-driven approach, illustrating how this approach differs from current decision-making processes incorporating a data-based approach. It investigates how the data-driven approach could be used in the legislative process and illuminates the challenges with this development. This article argues that fully reaping the benefits of the data-driven approach in a law-making context will require identifying and addressing the risks with this modern approach to decision-making in the modern society.
