The “beta testing” and development of demonstrators (using Aina Kit) of systems for monitoring, control and/or alignment in the use of AI/TL models and applications in Catalan.

We are looking for ideas for Beta testing and development of demonstrators (using Aina Kit) for monitoring, control, and alignment systems in the use of AI/LT models and applications in Catalan.

Eligible proposals must present solutions based on existing systems that have not yet used Aina Kit, as well as proposals for demonstrators without a pre-existing system, as long as they are operational within a maximum of six months from the contract signing.

The Beta testing and demonstrator development will focus on evaluating the process of monitoring, control, and/or alignment in the use of AI/LT models and applications, integrating Aina Kit resources as part of the solution, comparing the results with other options, and identifying potential issues in the provided technological resources, their integration into real-world environments (inference infrastructure, data integration, etc.), and/or the documentation and support provided.

Proposals are free to expand these basic areas and suggest anything they consider relevant for Beta testing and development of demonstrators (using Aina Kit) for monitoring, control, and/or alignment systems in the use of AI/LT models and applications in Catalan.

Some ideas that may fit within the Beta testing and demonstrator development for this challenge include:

• AI/LT governance systems in Catalan within an organization (e.g., related to organizational values, language style and tone, sector-specific legal restrictions, etc.).
• Systems to detect and mitigate malicious instructions, sensitive data, toxic and/or biased responses, hallucinations, prevention of restriction evasion attempts, etc.
• Systems enabling interpretable and explainable AI.

• Systems for compliance with applicable AI/LT regulations.
• Development of sandboxing environments to support AI/LT in Catalan.
• Cybersecurity mechanisms for the deployment of AI/LT in Catalan.

Proposals will be positively evaluated based on their social impact, their capacity to develop Beta testing and demonstrators in environments with a high volume of end users, and their ability to access data of sufficient quality and quantity, while strictly complying with current data protection regulations.

In this regard, proposals may include, along with the technical report, a commitment of collaboration in the development of the demonstrator by a company or end-user entity, specifying and explaining their participation.

Creativity and innovation are encouraged in the Beta testing activities, as well as in the development areas of demonstrators that, using Aina Kit resources, demonstrate monitoring, control, and/or alignment in the use of AI/LT models and applications incorporating Catalan.