Generative AI with Cybersecurity for Internet Applications development

Coach Name

Juan Juan

EU Organization

Novareckon SRL

Members

  • Alberto Stefanini
  • Alberta Bertin
  • Andrea Ceccaroni
  • Eleonora Dorissi
  • Christian Violi
  • Lorenzo Vandoni (Hal Service – end user)

CA Organization

Mind in a Box

Members

  • Jeremie Farret
  • François Monette

Project Overview

GALICIA (Generative AI with Cybersecurity for Internet Applications development) integrates generative AI with formal modeling to enable trustworthy, secure internet application development. It bridges AI-driven development (e.g. GPT-generated code) with formal verification (e.g. Coq, Prolog), focusing on safety, explainability, and reliability of software systems. Use cases included IoT apps, mathematical theorem formalization, and network protocol validation.

Methods and approaches

Hybrid Workflow

Prompting → Transformation → Verification → Feedback

LLM Integration

AI-generated source code from natural language

Formal Verification

Use of Coq, Prolog, Event-B, etc.

Agile, Human-in-the-Loop Design

PHP Laravel Framewor

with PostgreSQL, Docker, and Livewire

Open-Source Development

with GPL license

Key Achievements

Developed and released a complete working prototype

Verified real-world use cases (CRUD, TCP handshake, math proofs)

Held international workshops (Italy, Canada, Austria)

Published scientific paper at ASPAI 2025

Submitted paper to Automazione e Strumentazione

Released full open-source stack: GitHub Repo

Created 1 new job at HAL Service

Impact & Results

Scientific Impact

Advanced explainability and traceability in AI + formal code generation

Strategic Impact

Paved way for commercial models and domain-specific pilots

Economic/Industrial Impact

Validated usefulness for regulated sectors (IoT, telecom, healthcare)

Social Impact

Promotes secure and transparent AI-based software

Transatlantic Collaboration

Sustained cooperation and future research plans with Canadian partner

Publications and Open-Source Contributions

Future directions

Domain-specific pilots

in telecom, mobility, medical devices, and energy

Improvements

Semantic mapping for requirements Retrieval-Augmented Generation (RAG) Support for DSMLs Enhanced explainability & collaborative features

Commercialization Pathways

Evaluating spin-off potential

Community Engagement

Planning broader adoption via workshops, webinars, open-source forums