LGBTIQ communities and women, among the groups most affected by this inequality


The ITI technology center celebrates World Internet Day with advances in the European project ‘Aquitas’, whose objective is to guarantee that Artificial Intelligence (AI) is not only ethical and legal, but also equal.

The growing use of AI tools has raised concerns about the potential emergence of bias and discrimination. Along these lines, AI developers may have relied on historical data when creating the platforms, which could translate into injustices.

The initiative, funded by the European Commission, aims to develop an open environment so that developers and users who use Artificial Intelligence can experiment in a controlled way.

“To ensure fair AI that reflects European values ​​and is reliable, many AI tools that belong to public and private actors must be evaluated, repaired and, if necessary, replaced,” highlights, in a statement, Santiago Cáceres, manager of the project at the ITI.

However, the costs of replacing existing AI tools are high and unrealistic. Therefore, Aequitas will create an operational infrastructure that is capable of detecting, mitigating and repairing possible problems of injustice in AI tools, as well as defining new methodologies that avoid bias, so that new AI systems are developed in accordance with the principles fair by design.

The organization highlights that the notion of equity is also socially determined, and can differ between cultures and social groups and even within them. Along these lines, the Aequitas consortium has incorporated associations that represent various socially vulnerable groups and organizations with specific interests, in order to gather their points of view and experiences from multiple perspectives.

In this way, Aequitas will contribute to raising social awareness, especially among LGBTIQ communities and women, who are among the groups most affected by biases and injustice derived from AI. Likewise, it will target socioeconomically vulnerable groups who face a significant risk of digital inequality, such as children and the elderly.

According to Emilio Insfran, professor at the Universitat Politècnica de València (UPV) and director of the Software Engineering and Information Systems group (ISSI-UPV), “this project will provide informative content for the requalification and training of professionals in companies and public organizations “.

Specifically, the services developed through Aequitas will be available to the general public through the AI-on-demand platform. How this system works is that each tool, data set and use case can be tagged with a level of fairness.

In addition, the controlled experimentation platform will also be available as a stand-alone, downloadable tool for companies and public entities to explore various dimensions. This will allow them to test and repair their AI tools more effectively.

Coordinated by the University of Bologna, other Spanish entities such as the Polytechnic University of Valencia, the University of La Laguna and the Association for Economic and Social Development Hopscotch participate in the consortium.

Finally, the Aequitas project corresponds to file number 101070363 and is funded by the European Commission.