AIs are developing their skills at unimaginable speed, but can they pass the Turing Test? Created in 1950, the test is simple: an artificial intelligence has to convince people it is of human nature over a chat program. So far, no one succeeded, but scientists believe they have found a loophole #machinemagic
“Silence is a source of great strength”, said Lao Tzu. People can also see it as a response in itself, depending on context. That simple fact is the source of a theory a team of scientists from Coventry University in the UK put to the test. They looked at 6 transcripts form Turing tests and realized that when the machine stopped speaking, it made judges doubtful of the thing they’re talking to.
The pauses in conversation were simple robot malfunctions, but they were considered as much more by the innocent judges. When the judge is unsure, the machine has won. The thing is, can you really say the artificial intelligence is thinking by itself? No, it simply shuts up. It’s not a reason to celebrate as AI development goes, just an interesting finding in the way humans perceive them.
In this case, efficiency can’t be achieved by deceit.
Follow TechTheLead on Google News to get the news first.
