Descripción:

FAIRANDTRUST is an advanced research project dedicated to improving the fairness, transparency, and explainability of intelligent decision-support systems used in collaborative and multi-stakeholder environments. Modern organizations, governments, and institutions increasingly rely on artificial intelligence and data-driven systems to support complex decision-making processes. However, many existing intelligent systems still face important challenges related to bias, lack of transparency, unequal treatment of participants, and limited understanding of how decisions are generated. FAIRANDTRUST addresses these issues by developing new methodologies and computational models that integrate ethical principles directly into intelligent decision-making frameworks. The project focuses particularly on Group Decision-Making (GDM) scenarios, where several stakeholders or experts collaborate to reach collective decisions under uncertainty. In such contexts, fairness is not only about achieving equal outcomes, but also about ensuring that all participants and subgroups are treated equitably during the decision process. The project explores different fairness principles, studies their compatibility, and develops new fairness-aware models that consider protected attributes and consensus among participants. At the same time, FAIRANDTRUST emphasizes explainability, enabling users to clearly understand how and why decisions are produced by intelligent systems. Another important objective of the project is the development of explainable frameworks and tools that allow both experts and non-experts to interpret decision recommendations easily. By combining techniques from artificial intelligence, optimization, computational intelligence, and social welfare theory, the project aims to create trustworthy decision-support systems that are transparent, accountable, and human-centered. FAIRANDTRUST also includes the implementation of web-based platforms, simulation environments, and real-world intelligent decision-support applications. These tools will help organizations improve collaborative decision-making processes, reduce conflicts, and enhance trust in AI-assisted systems. Ultimately, the project contributes to the development of ethical and socially responsible AI technologies aligned with European values and trustworthy AI principles.

Propósito principal y objetivos:

The main objective of FAIRANDTRUST is to develop intelligent decision-support systems that are fair, transparent, and explainable for collaborative and multi-stakeholder decision-making scenarios. The project aims to improve the trustworthiness of AI-based decision processes by integrating ethical principles such as fairness, accountability, and explainability into group decision-making models and tools. Through advanced computational methods and human-centered approaches, the project seeks to support more reliable and socially responsible decisions in complex real-world environments. The key objectives of the project are as follow: 1. Study fairness in group decision-making by analyzing different fairness principles and their compatibility in collaborative environments. 2. Develop new fairness-aware decision models that consider consensus, stakeholder diversity, and protected attributes. 3. Create explainable AI frameworks that make decision processes transparent and understandable for both experts and non-experts. 4. Design comparison and evaluation tools to assess fairness, explainability, and user perception of intelligent decision systems. 5. Implement and deploy real-world decision-support systems integrating fairness and explainability into practical applications and web-based platforms.

Más detalles:

Referencia: PID2024-161073NB-I00

Entidad: Ministerio de Ciencia, Innovación y Universidades

Financiación recibida: 75.375€

Fecha inicio: 01/09/2025

Fecha fin: 31/08/2028