Detecting and Modifying the Non-inclusive Content in Mobile Apps
In recent years, the growing interest in inclusive and bias-free communication by tech firms and the software industry as a whole have been carried with a particular emphasis on technical documentation.
This research project aims to explore the varying degree of non inclusive language, classify them into categories and derive key insights that provide transparency into the graveness of the problem in respect to the industry of open source software projects as a whole.
The focus is on detecting and modifying the non-inclusive content in app metadata, such as the violent content in game apps, the gender bias in some app descriptions.
As part of the project, we are currently investigating how to utilise large language model such as GPT3, ChatGPT to achieve the goal.
Project Lead
Suyu Ma
Project Team
Prof John Grundy