2. Super-intelligence becomes a concern when we develop robots who are far smarter than we are—not just in a narrow sense (such as in chess or Jeopardy!), but in a more general sense. To date, AI's intelligence is narrow—that is, computers exhibit super-intelligence in only one area. But we soon will be seeing machines that exhibit what is called artificial general intelligence (AGI). When that comes about, will AGI decide to control us? Would it even care about us?
Another huge question is what impact AGI will have on human employment. AI already has stepped into mundane jobs (think of assembly line robots), but will soon disemploy many more people. What will we do with the millions of people who will no longer have work, because they've been replaced with robots? Very soon we'll see even rather sophisticated human careers—doctors, lawyers—terminated. We aren't prepared to deal with that problem.
Not a lot of effort is currently being applied to the issue of super-intelligence. Most researchers are rushing into the future, unable or unwilling to deal with this concern. Government barely understands the basics, let alone is currently prepared to address this matter.
3. Consciousness brings up another big question for which we currently have no good responses. Today's AI machines are most likely not conscious. But when—as evermore complex computers are built—will computers become conscious, or even sentient? Will they ever do so? Scientists are still struggling to define consciousness in humans and other life forms, let alone machines.
Is consciousness a continuum or is there a line which consciousness does not cross? Is a newborn baby conscious? A monkey? A tree? A rock? These questions are impossible to answer as we don't even yet know how to measure consciousness.
If AI machines do become conscious—after they become complex enough—will they be able to subjectively experience something? Could they experience pleasure or pain? How would we know? And would robots deserve rights, if they become conscious?
There are many unanswered questions about the future of AI. These machines may be a threat, or they may offer humans some type of salvation. Nobody knows at this point. Unfortunately, too few people—especially those in the public policy arena—are seriously considering these issues. Like so many technological innovations of the past, we may soon be dealing with huge problems, without having looked at the issues before they became problems.
No comments:
Post a Comment