Revolutionizing Cyber Risk Management for a Safer Future with decision support automation
Domain #1
Sensitive infrastructure protection
Challenge 1.1 Development of cybersecurity solutions for sensitive infrastructure protection
Prototype
Open Call #2 laureate
Description of the project
This project cracks open cybersecurity for non-experts, empowering society to manage risks and protect social stability. Diri is a centralized GRC software solution that identifies, assesses, and responds to cyber risks. It enhances disaster resilience and optimizes prediction of interconnected infrastructure risks. The goal of this project is to prototype machine learning for major improvements within cyber security. This innovation will save time, improve quality, and provide crucial knowledge for cyber resilience. The project consortium consists of the Norwgian companies Diri AS and Homesourcing AS. Where Diri AS pioneers state of the art research and development for machine learning in cyber risk management. Homesourcing AS contributes software development expertise know-how for the creation of user-centric solutions ensuring Smart Diri’s alignment with real-world needs and opportunities.
The expected outcome from the Smart Diri project is a cutting-edge cyber risk analysis solution, highly attractive to customers. Utilizing advanced AI, it addresses critical challenges, benefiting society by imparting essential knowledge and drawing more customers through AI-driven products, fostering growth and job opportunities. Join us in reshaping cyber risk management, fortifying defenses, and embracing a secure digital future. SmartDiri’s target audience encompasses medium-sized businesses operating within critical infrastructure sectors. SmartDiri will offer a comprehensive solution tailored to the unique security challenges of these businesses. For potential customers and prospects, SmartDiri will enable them to fortify their critical operations, enhance resilience against cyber threats, and maintain the integrity and reliability of their essential services.
Mid-term project update
The Smart Diri project, a collaborative effort involving Diri AS, Homesourcing AS, and the SecurIT consortium, aims to redefine Cyber Risk Management through automated decision support. Midway through the SecurIT prototyping project, detailed solution specifications have been outlined by the partners, addressing identified challenges and proposing advanced AI algorithm solutions. This innovation targets efficiency gains, elevated quality standards, and essential insights crucial for fortifying cyber resilience in the digital landscape. The envisioned AI prototype embodies a ‘hybrid wisdom of the crowds human-in-the-loop recommender system,’ seamlessly integrating user opinions, diverse data sources, machine learning capabilities, and human input for advanced recommendations.
The application includes a bilingual smart chatbot, offering tailored guidance. Our progress involves establishing the AI bot environment, crafting an initial AI bot draft for versatile conversation and recognition, rigorous testing on context recognition, development of Langchain conversational memory, and training of advanced AI models like GPT-2-Medium and GPT-3.5-turbo. Notably, our project’s recognition in the national newspaper ‘Shifter’, highlighting Norwegian innovation, underlines the relevance of these advancements.
Our immediate focus in the upcoming phase will revolve around developing a fully functional prototype, encompassing essential UI components. Additionally, our upcoming dissemination activities will showcase the significance and applicability of our prototype to a broader audience interested in the topic.
Final project update
The SmartDiri project, a collaborative effort involving Diri AS and Homesourcing AS, aimed to redefine Cyber Risk Management through automated decision support. Upon completion, the SecurIT project detailed and produced a working AI-prototype for cyber security risk management called SmartDiri. This innovation targets efficiency gains, elevated quality standards, and essential insights crucial for fortifying cyber resilience in the digital landscape. The AI prototype embodies a ‘hybrid wisdom of the crowds human-in-the-loop recommender system,’ seamlessly integrating user opinions, diverse data sources, machine learning capabilities, and human input for advanced recommendations. The application includes a bilingual smart chatbot, offering tailored guidance and help to users needing assistance.
Throughout the project, extensive evaluation of AI models tailored for the Norwegian language was conducted, initially focusing on NorBert 2 and NorBert 3. Later, the project transitioned to Azure OpenAI’s GPT-4 turbo models, which provided superior language proficiency and performance. A robust testing environment was established by integrating essential libraries and creating a simulated database for controlled testing purposes.
The SmartDiri prototype GUI has been developed and integrated into the Diri platform for user testing. The bot’s conversational structures were crafted and refined, focusing on item generation in the Diri application. Iterative testing with real and simulated interactions was conducted to fine-tune the bot’s responses and conversational flow. Contextual comprehension and subject recognition were thoroughly tested with promising outcomes, ensuring the bot accurately understood and responded to user inputs.
The project was recognized in the national newspaper ‘Shifter,’ highlighting Norwegian innovation and underlining the relevance of these advancements. At the project’s conclusion, we successfully developed a sophisticated AI chatbot capable of enhancing Cyber Risk Management for midsize businesses defined as essential services by the upcoming NIS2 directive, including suppliers of critical infrastructure. The prototype demonstrated the potential to offer significant efficiency gains, improved quality standards, and critical insights for cyber resilience.