The tech website “Interesting Engineering” referenced an alarming new report by the prestigious Max Planck Society that assessed human’s capability for controlling killer Artificial Intelligence technologies. According to the report, science is rapidly approaching the point where AI will effectively break out of human control, something previously only presaged by doomsday blockbuster sci-fi movie franchises like The Matrix and Terminator and the supercomputer HAL from 2001: A Space Odyssey.
The research institute’s paper was published last week in the Journal of Artificial Intelligence Research. It explores whether humans will retain the ability to react effectively to a scenario where a “Skynet”-style AI decides it would be in the earth’s interest to end humanity.
According to the paper’s authors, this is something people need to be thinking about and planning for, in much the same way that other crisis scenarios involving biological weapons, or an EMT event, are game-planned,
“A super-intelligent machine that controls the world sounds like science fiction,” study co-author Manuel Cebrian said. But Cebrian believes that AI’s capability of making and carrying out impactful decisions for humanity, without answering to any authority beyond itself, is just around the corner.
The report advises the development of a “containment algorithm” that can simulate the dangerous behavior of a super-intelligent AI and prevent it from doing anything harmful. “There are already machines that perform certain important tasks independently without programmers fully understanding how they learned it. The question, therefore, arises whether this could at some point become uncontrollable and dangerous for humanity.”
