# VALAWAI

- URL: https://giubots.net/research/valawai/

Most of my work as a PhD student revolves around the [VALAWAI](https://valawai.eu) project, funded by the European Union's Innovation Council.
It focuses on developing VALue-AWare Artificial Intelligence systems that integrate human values into their functioning.
Its goal is to enhance the ethical alignment and trustworthiness of AI applications by ensuring that they operate in ways that reflect the preferences of the people involved.

Within this project, my research focuses on social robots, with a particular interest in those deployed in a home.
How can we ensure that robots abide by our values and support positive human interactions?
How can we prevent manipulation and deception by robots?
How do we ensure that they act in ways that respect human autonomy and dignity?

## [Users' Perspectives on Value Awareness in Social Robots](https://giubots.net/publications/users-perspectives-value/index.md)

The first step of this line of search was a focus group that explored the expectations towards value aware robots of users. In total, 17 Italians between 15 and 29 years of age took part to a series of activities, such as brainstorming and drawing. The participants had no prior experience with robots, and the results revealed that they primarily viewed home-based robots as tools for automating household tasks rather than companions. Two major concerns surfaced: safety and privacy, reflecting media tropes and recurring themes in human-robot interaction discussions.

## [Values in Social Robots: Implementing Inclusive, Value-Aware Human-Robot Interactions](https://giubots.net/publications/values-social-robots/index.md)

An overview of my research so far, starting with the focus group, the work on LLM’s alignment and VLM’s alignment, embodied awareness and using VLMs for improved contextual awareness.

## [If They Disagree, Will You Conform? Exploring the Role of Robots' Value Awareness in a Decision-Making Task](https://giubots.net/publications/if-they-disagree/index.md)

Two identical robots, one “value-aware” and the other not, and one human. A set of images to be labelled. The humans take longer to confirm their choices when both robots disagree with them. This is just one of the findings of this study on value awareness in group settings by Giulia Pusceddu; you can read more about it in her blog post.

## Related Research Lines

This research grows along two main axes: what are values, and what does awareness mean.
The first challenge is to understand **what do we mean by values** in the context of human-robot interaction, an interesting topic explored [in this work](research/hri-values).
The next logical step is to explore how we can operationalise these values in robots, specifically through generative AI, in [a series of studies reported here](research/values-generative-ai).

The second challenge is defining **what is computational awareness**, how we can make robots aware of values and how they can adapt their behaviour accordingly.
While computational awareness can be applied to any algorithm, [in this research](research/embodied-awareness), I focus on those that have an embodiment, such as robots.