In the ever-evolving landscape of Artificial Intelligence (AI), the innovative potential is unquestionable. However, AI tools have the capacity to reflect, amplify, and even generate unfair biases, leading to exclusion, segregation, inequality, and harm, particularly affecting minority groups. Recognizing this challenge, the AEQUITAS project emerges as a research and innovation initiative aimed at addressing and tackling the multifaceted issue of bias and unfairness in AI across Europe.

The AEQUITAS consortium comprises a diverse range of partners, including experts from the High-Level Expert Group, universities, research centers, companies, SMEs, associations, and think-tanks. Their collective expertise ensures the project’s success in addressing the challenges of AI fairness comprehensively.

AEQUITAS is poised to make a significant impact on the AI landscape, fostering fairness, inclusivity, and trustworthiness in AI systems across Europe, while also serving as a global reference point for ethical AI development and deployment.

A Multidimensional Approach

AEQUITAS acknowledges that the problem of AI bias stems from various sources: the data used for training, testing, and validation, the design and development of AI algorithms, and the interpretation conventions encompassing AI systems. With this comprehensive perspective, AEQUITAS strives to tackle bias and unfairness on all these fronts.

European AI Ecosystem of Excellence and Trust

Aligned with Europe’s broader strategy of establishing an AI Ecosystem of Excellence and Trust, AEQUITAS places fairness as a central ethical principle, closely linked with others like respect for human autonomy, prevention of harm, and explainability. These principles are integral to creating AI systems that adhere to fundamental rights, laws, and governing principles, including the European Pillar of Social Rights, which champions gender equality, equal opportunities, and access to essential services.

AEQUITAS’ Threefold Mission

The AEQUITAS project aims to achieve three critical objectives:

• Technical Advancement: It seeks to ease the assessment, design, and development of fair AI systems, promoting the adoption of ethical principles, and offering a concrete framework for creating fair AI. The project will also develop novel algorithms for bias detection and mitigation and generate synthetic datasets for experimentation;
• Societal Impact: AEQUITAS strives to raise awareness about AI bias among marginalized groups, including LGBTQI individuals, women, socio-economically vulnerable populations, children, and the elderly. It provides training content to reskill and upskill the workforce in both public and private sectors;
• Inclusivity and Collaboration: Recognizing that the notion of fairness can vary among different cultures and social groups, AEQUITAS actively involves socially vulnerable and special interest organizations in its consortium. It aims to gather input and expertise from diverse perspectives on AI fairness.

AEQUITAS’ Deliverables

The project promises significant advancements, including:

• A controlled experimentation environment for assessing, mitigating, and repairing AI bias;
• Methodologies and guidelines for fair-by-design AI systems;
• Novel techniques for bias detection and mitigation;
• Synthetic datasets for experimentation;
• Real-world use cases in healthcare, human resources, and essential services;
• A service integrated into the AI-on-demand platform, making fairness assessments broadly accessible;
• Awareness campaigns and training initiatives;
• Strong Consortium for a Complex Project.