Artificial intelligence writes a book
Artificial intelligence can beat the Turing test by pleading the Fifth amendment and saying nothing.iStock

The Turing test, developed by legendary computer scientist Alan Turing and used to test the artificial intelligence of computers, has a major flaw.

A new study, published in the Journal of Experimental and Theoretical Artificial Intelligence, points out that the test, which was devised in the 1950s, could be successfully passed if the computer pleaded the Fifth Amendment and remained silent.

Authors Kevin Warwick and Huma Shah from Coventry University argue that a machine could plausibly pass the test by saying very little, or nothing at all. Previous attempts at defeating the Turing test have seen computers pretend to be children, but no matter how great their general knowledge, they are often unable to convince a human interacting with them (over typed messages) that they are also a human.

A critical point raised by the study is how the Turing test is based around a machine being discovered as a human based on what it does wrong, rather than what it does right. Its authors argue that a machine could know it isn't smart enough to act as a human, so pleads the Fifth Amendment as a way to hide its inabilities.

Warwick said: "This begs the question, what exactly does it mean to pass the Turing test? Turing introduced his imitation game as a replacement for the question 'Can machines think?' and the end conclusion of this is that if an entity passes the test then we have to regard it as a thinking entity."

However, Warwick argues, if artificial intelligence can pass the test by remaining silent, just as a human could choose to do in the same situation, this "cannot be seen as an indication it is a thinking entity, otherwise objects such as stones or rocks...could pass the test."

It isn't possible, the authors argue, for a human to determine whether or not they are talking to another human or a computer (or a stone wall) if they receive no reply. It could be a wall, or could be a human choosing to say nothing, or it could be a computer pretending to be a human by pleading the Fifth and saying no more.

To conclude, Warwick says that 'taking the Fifth' "fleshes out a serious flaw in the Turing test."