What's the Latest Development?
Computer scientists at the University of Louisville, Kentucky, are working on solutions to contain potentially dangerous artificial intelligence. Professor Roman Yampolskiy warns about unforeseen problems which advanced AI could cause: "It can discover new attack pathways, launch sophisticated social-engineering attacks and re-use existing hardware components in unforeseen ways." Yampolskiy has proposed trapping AI inside virtual machines that run from within a computer's operating system, limiting the access the AI has to essential software.
What's the Big Idea?
Should computers acquire what we recognize as human consciousness, they would be free to manipulate, seduce, bribe and blackmail humans to gain their freedom, says Yampolskiy. To prevent the situation from arising, he suggests making human safety a priority when using AI, allowing it to respond only in a multiple-choice fashion to solve specific technical or scientific problems. "If such software manages to self-improve to levels significantly beyond human-level intelligence, the type of damage it can do is truly beyond our ability to predict or fully comprehend," Yampolskiy said.
Photo credit: shutterstock.com