Ziyad Almahdawi
Cybersecurity Engineer
Security Matterz
Ziyad Al-Mahdawi is a Cybersecurity Engineer specializing in large-scale defense, SOC operations, digital forensics, and threat research. He explores practical strategies in AI security to strengthen systems against emerging adversarial techniques.
Speaker sessions
Data Integrity in AI: Defending Against Supply Chain Attacks in Model Training
The world is moving fast towards widespread adoption of AI model training across industries, transforming how businesses and governments operate. Yet, as AI systems become a significant part of our critical decisions, their security hinges on one fragile point: the integrity of the data that fuels them.In May 2025, the NSA’s Artificial Intelligence Security Center, together with CISA, FBI, Australia’s ACSC, the UK’s NCSC, and New Zealand’s NCSC, released AI Data Security: Best Practices for Securing Data Used to Train & Operate AI Systems. This first-of-its-kind joint guidance tackles the weakest link in AI deployments — the integrity of the data powering them.This session will break down the three primary risks — data supply chain compromises, maliciously poisoned data, and data drift — and show how these can be exploited by adversaries to subvert AI models. Attendees will learn practical, field-tested measures for securing AI data lifecycles, from provenance tracking and digital signatures to continuous validation pipelines. We’ll examine real-world threat scenarios, explore the implications for AI-enabled SOCs, and outline a blueprint for operationalizing these recommendations. By the end, you’ll know how to harden AI systems before attackers can turn your data into their backdoor.
- 15:40
- Wed
- 03 Dec
Stage:
Briefings 1
Sessions Type:
Presentation