The artificial intelligence (AI) revolution raises questions that used to exist only in science fiction. Will AI replace artists and writers? Is it any good at diagnosing diseases? What if it decides to deploy nuclear weapons?
Some researchers have bypassed these relatively pressing questions with some slightly more abstract concerns. If humans succeed at developing artificial intelligence that truly thinks like humans, they ask, will it have feelings or the capacity to suffer? According to Forbes, negative emotions like fear or anxiety could impact how it completes certain tasks and processes. For example, what if an AI system designed to make stock trades started panicking every time the stock market dipped? Could an AI-powered self-driving system develop road rage?
It sounds ridiculous, but they may have a point. According to the journal Nature, the implications of an artificial consciousness are too serious to totally ignore, and humans should figure out how to identify it in the (possibly very unlikely) event that it really happens. And in that case, there are still more questions to answer, like how humans should treat a sentient machine and whether it deserves rights.
Researchers aren’t hoping to find emotions that are unlikely to exist — instead, they want to develop a process to know for sure that they don’t, so that humans don’t devote time and resources to caring for machines instead of each other.