Only 16.4% of Italian businesses use AI effectively. Discover how to join them.

Download the whitepaper

AIA Guard

Artificial Intelligence Attack Guard

Goal

Release AIA Guard, a TRL9 cybersecurity solution specifically designed against Artificial Intelligence Attacks.

Use cases

1

Measure the AI model drift due to data poisoning.

2

Dataset sanitisation and data privacy preservation.

3

Evaluate the AI application trustworthiness and explainability.

Research programme

SecurIT - FundingBox (cascade funding)

Timeframe and status

2023 - 2024 (Completed)

Topic

Cybersecurity

Project website

Datrix Roles and Contributions

1 – Consolidate in a single and user-friendly platform.

2 – Release NLP-driven dataset sanitisation module.

3 – Release user-friendly module for data poisoning and AI model hardening.

4 – Release advanced module for AI explainability.