The language and length of privacy policies make them impractical to be read/understood by common people, this study aims to change that. We will target from sensors and wearables to health-related IoT hubs. After identifying relevant devices, privacy policies will be downloaded from their online websites and a dataset will be created.
This study will try to provide a 3 pronged analysis of privacy policies within these devices through the use of:
- Topic modeling
- A pre-trained privacy policy classifier
- Large language models (LLMs) trained on privacy policies/legal documents etc., to (a) summarise the text and (b) identify contents in existing privacy policies
Required/ Recommended Qualification
The ideal candidate should have knowledge of text-based machine learning tools (Large Language Models, Topic Modelling, etc.) and an interest in privacy and society.
Contact Persons
• Rana Tallal Javed (ranaj@ifi.uio.no)
• Thomas Plagemann (plagemann@ifi.uio.no)
Parrot: This research project is associated with the Parrot project (Privacy Engineering for Real-Time Analytics in Human-Centered Internet of Things).
• Project Site: [UiO - Institutt for informatikk. (2020, June 24). Parrot: Privacy Engineering for Real-Time Analytics in Human-Centered Internet of Things. https://www.mn.uio.no/ifi/english/research/projects/parrot/index.html (Retrieved 7 September 2023)]
• Vision of Parrot: [Plagemann, T., Goebel, V., Hollick, M., & Koldehofe, B. (2022). Towards Privacy Engineering for Real-Time Analytics in the Human-Centered Internet of Things (arXiv:2210.16352). arXiv. http://arxiv.org/abs/2210.16352 (Retrieved 7 September 2023)]