The Trust aWare partnership is working on Android devices
When you install an app on a device, you are provided with information on the functions supported by the programme, such as access to your camera, contacts and location. Each authorisation is accompanied by a description explaining what can be used and for what purpose. Yet, how many of us click on “agree” without reading this information? Do we really know what we are agreeing to? These are usually long, boring and often deliberately confusing texts written in a formal and legalistic style that are hard to read and understand.
Project such as Trust aWare help us to understand if we are acting superficially and what we can do to protect our personal information. And this is no easy feat considering that the average user would require more than 70 full-time working days to read the privacy policies of the services used in one year!
If you understand English, we recommend you read this interesting article published on the project website that describes one the first actions concerning the analysis of texts written in natural language on the use of software.
In the context of the Trust aWare Project, partnership experts are implementing a framework of notes for privacy policy and app descriptions for apps on Android devices by applying cutting-edge techniques for the analysis of natural language and deep learning. The objective is to improve the understanding of documents and promote privacy in the use of apps, so to satisfy user expectations.
One of the key challenges is to help final users to better understand the information provided when they install and use apps – Android, in this case – to evaluate if the requested authorisations are truly necessary for their functionality.
In order to efficiently achieve this objective, accessible and simple-to-use models are being developed. In the coming months, an initial graphic interface will be available to help users comprehend the documentation through an interactive visualisation. Users will be able to select the desired category (i.e., personal information) and visualise only the sections pertinent to their interest in the documentation.
