AIAGENT4CYBER – Multi-Agent AI-Based Cyber Defence

Project Description:

AIAGENT4CYBER is a European cybersecurity project coordinated by ITCL that proposes an innovative approach based on multi-agent artificial intelligence for real-time threat monitoring, adaptive intrusion detection and automated incident response.

The system relies on specialised AI agents that collaborate with one another, learn continuously and adapt to new and evolving threats, enabling proactive and resilient defence in complex digital environments.

General Objective:

To develop and validate an advanced cybersecurity system based on multi-agent artificial intelligence, capable of autonomously, scalably and reliably monitoring, detecting, anticipating and mitigating cyber threats, reducing response times, human intervention and false positives in complex digital environments.

Specific Objectives:

Design and implement a multi-agent AI framework for the continuous monitoring of cyber environments, with the ability to autonomously detect anomalies and known and emerging threats.
Develop an autonomous AI-based Intrusion Detection System (IDS) capable of analysing network traffic in real time, classifying attacks—including previously unknown ones—and generating immediate, highly accurate alerts.
Implement automated response and mitigation mechanisms capable of blocking malicious traffic, isolating compromised systems and deploying countermeasures without human intervention, while minimising the impact on legitimate users.
Ensure the scalability, modularity and adaptability of the solution for deployment across different sectors, organisation sizes and technological infrastructures.
Improve collaboration and coordination among AI agents by enabling efficient communication, shared knowledge, distributed decision-making and trust mechanisms.
Ensure ethical, secure and compliant deployment, guaranteeing privacy, data protection, transparency, fairness and regulatory compliance, including GDPR.
Validate the system through real-world use cases, demonstrating its robustness, performance, adaptability and operational effectiveness under real-life conditions.

Challenges Addressed:

The project addresses several structural limitations of traditional cybersecurity approaches:

The inability of conventional systems to adapt to new and unknown threats.
Excessive reliance on human intervention, which slows down incident response.
High false-positive rates, which reduce operational efficiency.
Limited scalability and flexibility in heterogeneous and distributed environments.
A predominantly reactive approach, acting only after damage has already occurred.

AIAGENT4CYBER addresses these challenges through an autonomous, preventive and collaborative approach capable of anticipating attacks, reducing the impact of incidents and improving the digital resilience of organisations, particularly SMEs.