As smartphone use increases dramatically, so do studies about the negative impact of overusing technology. Nowadays, several tools for monitoring and controlling device usage and achieving “digital wellbeing” exist as off-the-shelf products. These Digital Self-Control Tools (DSCTs) allow users to track their smartphone usage and to define interventions, e.g., timers and lock-out mechanisms, to self-regulate their behavior with digital devices. Unfortunately, several studies demonstrated that existing DSCTs tools are often not effective, especially in the long-term. To promote a more conscious use of technology, prior works in the digital wellbeing context agree on the importance of learning how to properly use technology: instead of blocking a “bad” user’s behavior, for instance, a DSCT could be used as a learning support, e.g., by suggesting desirable alternatives, or it might help the user to think about the negative aspects of her choice.

This thesis aims at exploring novel conversational DSCTs that proactively support users in learning how to properly use their smartphones according to their specific needs. The following are planned:

  1. Identification of a set of strategies that should be included in a conversational DSCT for smartphones. This entails the identification of what the user can communicate to the application, e.g., her initial preferences, and what the tool can “suggest” to the user. As an example:
    1. the user initially tells the DSCT that she would like to reduce the time spent on social media;
    2. the DSCT monitors the usage of social media, and detects that the user is spending a considerable amount of time on Facebook;
    3. the DSCT tool proactively starts a conversation with the user by explaining the detected problem and by suggesting different alternatives to mitigate the user’s behavior, e.g., setting up a timer or receiving motivational quotes.
  2. Design and implementation of a conversational DSCT with the inclusion of the identified strategies. The DSCT will be developed as an Android application, and it will exploit existing NLU platforms like DialogFlow or Rasa.
  3. Evaluation with users, either in the lab or in-the-wild, to assess the usability and efficacy of the developed tool.


Giulio Piacentini

Thesis Details

Luigi De Russis, Alberto Monge Roffarello
Master Degree in Computer Engineering