The talk will focus on AI and cybersecurity. Currently used by hackers for deep fakes, will it evolve into AI specialised in network attacks? Will there be AI on the defence side? Will we see a battle between the best AI (attack versus defence)? What protection mechanisms could prevent hackers from using AI or compromising AI? Are we in a dangerous transition phase where AI does not yet have sufficiently robust ethical mechanisms that cannot be circumvented? These points will be addressed from a general, non-technical perspective.