Autonomous systems — like driverless cars — perform tasks that previously could only be performed by humans. In a new article, artificial intelligence ethics experts argue that current safety regulations do not plan for these systems and are therefore ill-equipped to ensure that autonomous systems will perform safely and reliably.