Safe, ethical and secure: Robots you can rely on Professor John A McDermid OBE FREng, Director, Assuring Autonomy International Programme

Event details

Department of Computer Science Lecture

Computer systems and software have been used in safety-critical applications, where their behaviour could impact human health and wellbeing, for at least 50 years. These uses have evolved from relatively simple applications, e.g. syringe pumps used to deliver medicines, via sophisticated systems supporting humans, e.g. railway signalling and aircraft flying control systems, through to autonomous systems – where computers and software control physical systems without human intervention. Such modern systems include robots in factories and those that can support the elderly in independent living. Often these systems include software technologies such as machine learning and computer vision. These technologies are not just used for robots – they can give clinical advice, they can drive trains and cars – the applications are almost endless. But, even with the current systems, there can be ethical concerns. Will a clinical advisory system treat all patients equally, or will it give advice better suited to one ethnic group than another? Will an autonomous road vehicle be more likely to hit and injure pedestrians than other cars? Will a breach in system security cause harm via any other system connected to it? The talk will illustrate the evolution of such critical systems over (at least) 40 years and consider what can be done to make them safe, ethical and secure – hence producing robots and other advanced systems you can rely on