At the workshop led by dr. Zoltán Gyurasz a comprehensive view of the connection between artificial intelligence and cybersecurity was heard - a topic that today can no longer be perceived only as the protection of systems, but rather as a question of their robustness, trustworthiness and legal framework.
👉 Key ideas:
• AI as a new security vector
Artificial intelligence systems face specific threats - from adversarial attacks (avoidance) to data poisoning to information leakage or component compromise. Protection must cover the entire ecosystem.
• Security ≠ just technology
Effective AI protection requires a combination of:
✔️ organizational measures
✔️ technical solutions
✔️ specific AI mechanisms
Standardization and an interdisciplinary approach also play a vital role.
• AI as a defense tool
Artificial intelligence is already strengthening Cybersecurity today - especially in the area of:
🔎 anomaly detection
⚙️ reaction automation
📊 attack analysis
🤝 information sharing
• AI as an attack tool
However, the same technologies also enable more sophisticated and convincing cyber attacks. This is a classic "dual-use" problem.
👉 Conclusion:
AI security cannot be narrowed down to a legal framework or isolated technical measures. The key is:
✔️ protect the system as a whole
✔️ ensure data quality
✔️ maintain transparency
✔️ implement human supervision
✔️ manage risks throughout the entire life cycle
🎥 You can watch the entire recording at: https://www.youtube.com/watch?v=SJycXESfcWE