Generative AI with Cybersecurity for Internet Applications development

Mind in a Box is very happy to launch project GALICIA, with its Italian partners Novareckon and HAL Service.

GALICIA is a project funded by the European Union, within the framework of the NGI Sargasso program.

The aim of the project is to test a novel approach to digital resilience verification by testing LLM generated code for correctness and security on a set of case studies, aiming to ensure compliance with user requirements and given standards. The ambition is to verify source code generated by Generative AI and analyze its limits, thus building trust in Generative AI. GALICIA aligns with the increasing demand for compliance in industrial automation and the need for fast and low cost software production.

Main expected results of GALICIA:

GALICIA will provide a platform for code verification on a set of test cases in automation, encompassing a large case study of industrial relevance. It will encompass a two-step verification of LLM generated code:

• Generation through LLM of source code, from user provided natural language functional requirements;
• Compliance verification of the generated code with users’ natural language security specifications through the NuSMV theorem prover and other formalisms.

The project’s duration is 9 months, and results will be posted on a soon to be setup website.