The Machines Are About To Become Self-Aware And They May Not Like Us Very Much

"Open the pod bay doors please, Hal." "I'm sorry Dave, I'm afraid I can't do that." So said the sentient computer HAL at the controls of the Discovery One spacecraft to astronaut Dave Bowman in the 1968 Stanley Kubrick classic, "2001: A Space Oddyssey." While computer-versus-human conflict has been foreseen in many films since then, a leading artificial intelligence researcher is now making the case that we need to start planning for the day that artificial intelligence combined with lethal capabilities will pose a real challenge to humanity. Roman V. Yampolskiy, a respected artificial intelligence researcher and the director of the Cybersecurity Laboratory at the University of Louisville, is the author of a new study, "Taxonomy of Pathways to Dangerous AI," due to be presented for the first time Saturday at the Association for the Advancement of Artificial Intelligence conference in Phoenix, Arizona. His paper is an attempt to spark a serious, intellectual discussion what controls humans can put on machines that don't exist yet.