Artificial intelligence is no longer just a useful tool for work. In their interactive demonstration, the Security@FIT research group will guide you through the risks associated with it. We will show you the latest developments in the field of deepfakes, such as the impact this synthetic content has on evidence in court proceedings. You’ll learn how deepfake attacks are evolving, but also, hand in hand with this, the defence methods designed to provide users with more information about suspicious media.
We won’t forget the fundamental questions either:
· How are deepfakes created?
· How similar is a synthetic voice copy to the original?
· How difficult is it to impersonate someone else?

As well as deepfakes, we’ll also be looking at the security risks posed by large language models (LLMs): you’ll learn just how easy it is today to alter any piece of knowledge within a language model, and we’ll guide you through the security implications of this technology.