Author: Angeliki Tsanta (EOS)
AI4CYBER is an EU-funded research project under Horizon Europe (GA ID: 101070450) which brings together 13 partners from 7 EU Member States to develop and test smart cybersecurity services leveraging AI and Big Data. Under the coordination of Tecnalia, the project is running for 36 months from September 2022 to August 2025 to deliver the AI4CYBER ecosystem framework which will bring together 11 AI-based services from different cybersecurity areas.
This blog explains the project’s added value for cybersecurity stakeholders, presenting its vision, solutions and next steps.
AI4CYBER Vision
AI4CYBER aims to enhance the security and resilience of critical systems against advanced and AI-powered cyberattacks. It plans to do so by advancing the existing state-of the art AI techniques to create specialised solutions for testing, detection and threat intelligence and integrate them in one framework. This framework will also be evaluated by end-users and adjusted to reflect their feedback.
The work undertaken by the consortium will benefit:
- Operators of critical systems that have been identified under the NIS2 Directive 2022/2555 as essential or important services and called to provide a high level of cybersecurity. AI4CYBER services will leverage AI to facilitate, automate and improve the efficiency of their work.
- Cybersecurity solution providers that can integrate the AI4CYBER solutions to their services and provide more competitive tools into the market.
- The scientific community that can use the research undertaken by the AI4CYBER consortium regarding advanced and AI-powered cyberattacks to progress its own work.
- Policymakers who can monitor how AI-based services can improve cybersecurity and protect critical infrastructure in practice.
AI4CYBER Solutions
As mentioned above, the consortium is developing 11 AI-based services that can be embedded in cybersecurity tools. These will be:
- AI-Driven Software Robustness and Security Testing: AI4VULN – Code testing and AI4FIX – Vulnerability fixing
- Smart Detection and Analysis of Advanced Attacks: AI4AICTI – Cyber Threat Intelligence improvement; AI4SIM – Threat Simulation; AI4FIDS – Federated Detection of threats; AI4TRIAGE – Incident triage
- Self-Healing and Autonomous Response: AI4SOAR – Security Orchestration, Automation Response; AI4DECEIVE – Deception and honeypots; AI4ADAPT – Long term adaptation; AI4COLLAB –Information sharing and collaboration
- Trustworthy AI for AI4CYBER solutions, including xAI, fairness and security of AI: TRUST4AI – Trustworthiness of AI
More information on each solution can be found under the Solutions tab in our website, while as the project progresses, more blogs, videos and articles explaining how each solution works will be published.
Validation
In addition to developing a new breed of smart cybersecurity services, AI4CYBER will gather end-user feedback to ensure that the project outputs reflect the needs and specifications of the industry. The consortium includes three operators of essential services in the energy (Public Power Corporation, PPC, Greece), banking (CaixaBank, CXB, Spain) and health (Hospital do Espírito Santo, HES, Portugal) sectors who are in constant exchange with the solution developers. Three use case demonstrations will thus take place to evaluate the effectiveness of the AI4CYBER methods.
Next steps
The consortium is currently developing the initial framework architecture of the AI4CYBER solution with the aim of finalising its design by August 2023 and moving to a first release in August 2024. More information about the research and upcoming results will be published in the coming months.