Ethical PROMISE NFR Dataset
Description
The Ethical PROMISE NFR Dataset is an expanded and enhanced version of the original PROMISE Non-Functional Requirements (NFR) dataset. Whereas existing datasets only consider software quality attributes from an operational perspective, such as security, performance, and usability. This particular version of the dataset includes a socio-technological taxonomy focused on contemporary AI ethics and algorithmic governance. The multi-label annotated dataset consists of 9 ethical concerns categories, Privacy Fairness Transparency Accountability Informed Consent Accessibility Sustainability Human Autonomy and Bias This dataset can be used to train, benchmark, and test sophisticated ML classifiers and Large Language Models (LLMs) to detect ethical issues at the requirements engineering stage of software development.
Files
Steps to reproduce
1.This dataset contains two labels one functional label and another ethical label. Functional label consisting of functional and non-functional requirements as specified in PROMISE NFR Dataset. Ethical label consists of 9 ethical categories. 2. Refer readme document to know details of dataset 3.This dataset can be used to train, benchmark, and test sophisticated ML classifiers and Large Language Models (LLMs) to detect ethical issues at the requirements engineering stage of software development.