Protecting Artificial Intelligence – A New Frontier in Cyber Defense
✨ Our very first research seminar with a speaker in Gatineau and a live broadcast in Montreal was a great success!
We had the privilege of welcoming Myria Bouhaddi, a computer science professor at UQO, who immersed us in the world of cyber defense in the age of artificial intelligence.
A big thank you to everyone who attended in Gatineau and Montreal for their interest and engagement in the discussions. Stay tuned for our upcoming events! 🚀
Summary:
Artificial Intelligence (AI) has become a key strategic tool for strengthening organizational cybersecurity. Its learning and adaptation capabilities make it a powerful asset for identifying emerging threats, detecting abnormal behaviors, and enhancing the resilience of digital infrastructures. However, AI itself is increasingly targeted by sophisticated attacks, threatening the confidentiality and integrity of the models it powers.
This seminar explored the main attacks targeting AI models, with a focus on inference attacks, which exploit information leaks to extract sensitive data. The research conducted at the Laboratory for Computer Security Research (LRSI) at UQO was also presented, particularly on the use of adversarial examples as a defensive mechanism to enhance model robustness.