When cybersecurity becomes cAIbersecurity
AI, and especially machine learning, is hot. With the advent of ChatGPT and image generators we have made a sudden unexpected jump in performance, which gave rise to fears that soon AI will take over the world. Do we really need to fear the coming of our artificial overlords? Or is it the use of AI that we need to worry about? Because, while AI is successfully employed to keep our networks safe, it is increasingly used by adversaries as an offensive tool. A cat and mouse game, with one difference: what if your AI itself gets attacked?
In this keynote Peter Zinn, former Innovation Liaison of The Hague Security Delta and Sr. High Tech Crime Advisor of KLPD (National Crime Squad of The Netherlands), will discuss the pAIn, gAIN, and drAIn of artificial intelligence. Some of his take-aways:
- The fears of AI taking over are unfounded. Yet. What we DO need to be afraid about is the users of AI.
- AI is increasingly and successfully used in cyber-attacks, in every stage.
- AI has been used in cybersecurity for a while. It’s being used by both sides and will lead to an arms race where the security team will need to reinvent their role.
- AI already falls victim to bias and variance. We need to be able to recognize it if our AI is compromised by an adversary.