The main goal of this project was to collaborate with indigenous languages interpreters to co-design a Conversational Agent (CA) which could aid them to collectively generate data that: 1) could help them make visible the numerous problems and barriers they face working as interpreters, particularly within Mexico's justice system and; 2) increase their agency and decision-making power on issues related to the interpretation, such as in public policy. In this sense, the project intended to address the existing need to map the current state of indigenous languages interpretation in Mexico, to provide data that, for example, could help assess existing public policies related to interpretation and access to justice for indigenous Nations, or that could help identify indigenous languages and variants of such languages in need of certified interpreters.
Central to the project was the development of a methodology to align the system and the research process, in all its stages, with the principles promoted by the project, which are: gender perspective, co-design, shared benefits, digital autonomy and data sovereignty. Important elements of this methodology included a series of workshops with indigenous language interpreters, the creation of a Research Protocol and Collaboration Agreements and the development of strategies to incorporate the CARE Principles for indigenous data management into the project.
In this context, some of the research questions that motivated the work include:
How to develop co-design mechanisms to enable the inclusion of the end-users in all stages of the life cycle of an AI system?
How to align technological development with the principles of gender perspective, collaboration, co-design, digital autonomy
and data sovereignty?
Which factors are important to assess the feasibility of developing an AI solution based on the aforementioned principles?
How to incorporate Feminist theory and Feminist practices into AI development?
