Dr Madeline Cheah, Associate Technology Director, Cambridge Consultants
What happens when your system starts to think and do things for itself? Autonomous systems don’t just follow instructions; they act, adapt and deceive. As we hand over more agency to machines, we not only have to deal with the traditional attack landscape, but open the door to novel threats, from subtle manipulations to emergent misbehaviours to weaponisation by adversaries. This talk unpacks the risks of giving AI agency, explores how things can go sideways fast, and offers strategies to keep these systems assured.