About the project
Interested in building trustworthy and responsible AI? Then join us at the University of Southampton to incorporate ethics into AI and build models and systems which can act in alignment with the ethical preferences of their users. Test and deploy these across diverse applications like transportation, energy management, disaster response.
A variety of AI based applications like chatbots and recommendation algorithms are designed and deployed for the citizen end-users who may not be aware of AI advancements. And in many cases these applications make decisions on behalf of their users, for example an autonomous vehicle driving or and smart home managing the energy consumption at home. In such cases, it is crucial to ensure that these AI-based applications are ethically aligned with the preferences of the end-users.
In this project, you will focus on Ethical Alignment in Citizen-Centric AI which involves:
- to computationally model different ethical theories
- to develop methods for eliciting the ethical preferences of the end-users
- to develop techniques for aggregating the ethical preferences of multiple users
- to ensure that the AI based application aligns with the aggregated ethical preferences.
For this research, you will use techniques such as mathematical modelling, agent-based simulation, and game theory.
In the current times of rapid AI advancements, you will build frameworks to ensure that AI systems act in alignment with the ethical preferences of their users and can be applied across diverse domains like transportation, energy management, and disaster response.
If you are familiar with mathematical modeling, have some programming experience and are interested in the ethics of AI, then do apply! You will be part of the Citizen-Centric Artificial Intelligence Systems (CCAIS) team in Agents, Interaction and Complexity (AIC) research group.
For further details about this research direction, see the following paper: https://eprints.soton.ac.uk/493529/1/paper_473.pdf
