Exploring project ideas

>Design an intelligent bot that helps users with privacy matters; like configuring their settings and answering questions about privacy policy/terms of service. Designed in such a way that minimal amount of user information is needed.

(adaptive interactive system for privacy domain)

Test different designs/levels of autonomy in configuring the settings, how does this affect user satisfaction/trust > which best approach

>How collect domain specific data from users to help train chatbot in a way that preserves privacy, is engaging and is tagged/processed/prepared automatically (as far as possible) > what possible / best approach

>Bot that teaches people about digital rights and current privacy problems + potential solutions

>Design and implement a privacy preserving chatbot, based on more complicated machine learning models. The bot can be general purpose, not specifically as how would be implemented on iYYU. (possibly the teaching bot).

Some works where something similar is done / described:

Privacy preserving chatbot conversations - biwas et al.

Privacy preserving dialogue system based on argumentation - fazzinga et al.

>The AI privacy toolkit by IBM, about data anonymisation and minimalisation. This toolkit could serve as a great basis for research questions about its efficacy in certain domains or shortcomings for example. Specifically applying this toolkit to a chatbot use case, one that makes use of natural language processing. (An accompanying paper about the anonymisation techniques used can be found here, and for minimisation here)