We refer to empowerment as the degree of autonomy and self-determination in people and communities. We study how AI can help individuals become more self-sufficient and advance in their goals. The objective is to help people become more confident in controlling their lives and enhance their potential to influence the world around them.

On the algorithmic approach, our work expands on the knowledge of AI empowerment that explores the potential an agent perceives that it has to influence its environment. We explore how different machine-learning approaches can help agents identify and learn from human interactions and adapt according to the immediate benefit for humans. Evaluating the AI outcomes is a crucial component of our technology to promote Trustworthy AI solutions.


We study how AI can intervene in our daily lives to promote wellness in different dimensions, such as mental, physical, economic, and emotional well-being. We focus on understanding how to measure and parametrize well-being measurements that algorithms can use to optimize their outcomes. Our group explores how machine learning models that learn from subjective factors can gradually improve their perceived performance.

Our approach starts with applying the quantified-self method, where users are willing to understand how they perform periodically. Humans can assist algorithms in understanding the meaning of the collected data and how these digital traces can be linked to a direct benefit for them. We develop technology to track these digital traces to train algorithms to help individuals achieve their goals.


  • Smart workplace tooling can increase workers' wages and satisfaction (performance)
  • Algorithmic unions can combat unfair digital practices of platforms (justice)
  • Smart upskilling can provide a direct learning path to achieve a goal (education)
  • Smart interview trainers can prepare candidates to get a job position (recruitment)
  • Smart mentorship can guide novices and connect them with human mentors (mentoring)

Knowledge Base

  • We evaluate our tools in real-world settings to understand the real impact of AI-powered tools.
  • Our solutions collect constant feedback from users and adapt to users evolving preferences.
  • The main goal is to learn from users' goals and help them to achieve them.
  • Anonymized datasets are shared to help the community to analyze performance data.