LAWRIE McFARLANE

Smile, bot! How cleverly programmed robots could attain personhood

In the Star Trek episode The Measure of a Man, a scientist wants to disassemble Commander Data to see how his innards work.

Following a hearing, the scientist is sent packing, and Data is declared sentient, with all the rights of a human being. Was this the correct decision?

Alan Turing would certainly have thought so. Readers who’ve seen the movie The Imitation Game will remember the Cambridge mathematician who played a leading role in cracking Germany’s Enigma code.

Turing also built the world’s first computer, and later came up with a way to decide when a computer could be called intelligent. His method — the Turing test — was to place a computer in a sealed room and connect it to a teletype machine.

If humans communicating via the teletype couldn’t tell whether they were talking to a machine or a person, then the computer should be deemed intelligent.

So let’s say someone comes up with a robot (for that’s what Data really is) that observers are unable to distinguish from the real thing. I suspect that day lies many years in the future.

Yet if such a robot were produced, would it then, via the Turing test, be sentient? That is, would we be obliged to treat it as if it were a real person? If Turing (and Star Trek) are right, that might seem the inevitable conclusion.

I’ll return to that viewpoint in a moment, but first a counter argument. Let’s say I own da Vinci’s Mona Lisa — unlikely, since it’s in the Louvre — but let’s suppose I do. You, on the other hand, have a fake version so cleverly produced that no expert can tell one from the other.

Is your painting then genuine? Clearly not. It might be indistinguishable from the real one, but da Vinci didn’t paint it.

In these, and countless other instances, we ignore degrees of similarity, even if the match is exact, because something more fundamental motivates us. That something is the item’s history and point of origin.

Da Vinci’s Mona Lisa is genuine, because da Vinci himself painted it. We are admiring a work of art that connects us to an era 500 years in the past.

On this way of looking at things, Data is not sentient, and neither is he entitled to be treated as a person. He (or more accurately it) is merely an artful reproduction. Put another way, just because something walks like a duck and quacks like a duck doesn’t make it a duck.

So which viewpoint is correct? My guess is that it all depends upon the subject matter.

When it comes to inanimate objects such as paintings, we care very much about authenticity. For collectors of antiques, pedigree and provenance are everything.

However, when dealing with a machine that possesses the outward appearance of human qualities, a different set of values comes into play.

For most of human history, we knew very little about what made us sentient. Aristotle thought the heart was the seat of intelligence and the brain was merely a cooling device.

Rather than internal anatomy, the features that draw one human to another are the kind of interactions that we recognize instinctively as human. They include facial gestures, voice tone, emotional responses, caring behaviour and so on.

But behaviour and appearance are reproducible in a robot by clever programming. Not today, perhaps, but sooner or later.

If our species spends several decades with such beguiling companions, I’m willing to bet we end up granting them personhood. Their human characteristics will simply be too compelling for us to treat them as mere machines.

The moral and legal quandaries that come with this are formidable, however. For we will be boldly going where no one has gone before.