'They already face the same moral dilemmas as human beings – whether to profit at the expense of others, or whether the common good wins over selfishness.'

Philip Purser-Hallard asks if science can give a human being the power of a God, what kind of God will they become?

Last year I promised that a future column would look at the spiritual status of artificial intelligence (AI) – the so-far theoretical possibility of computers (or their software) with thinking, self-aware personalities.

Given the recent research showing that experimental robots designed to evolve and communicate grow up to be capable of lying, murder and heroic self-sacrifice, this seems as good a time as any.

These robots’ intelligence is rudimentary at best, but they already face the same moral dilemmas as human beings – whether to profit at the expense of others, or whether the common good wins over selfishness.

If and when software becomes capable of human levels of cognition and emotion, we’ll face this dilemma ourselves. How should we as a species treat our artificial progeny?

Will we demand their love and loyalty even as we set them to work for our own ends, or will we allow them to define their own rights and responsibilities?

Will they, like human beings, be capable of sin and salvation? Will they have souls? How does the Good Shepherd feel about electric sheep?

The Modern Pygmalion

Issues like this – or more accurately, the various things science fiction (SF) writers have said about them – are what I wrote my doctoral thesis about, so I’ve given them a fair bit of thought.

The fundamental question is: if science can give a human being the power of a God, what kind of God will they become?

Although the story is as old as Galatea and the Golem, it was first expressed as SF in Mary Shelley’s genre-founding novel Frankenstein.

There Frankenstein’s creature, after reading Milton’s Paradise Lost, complains of his creator’s inhumane treatment – rejecting Adam’s obedient role in creation and instead taking that of the resentful angel Satan.

Since then, the figure of the creature in SF has taken a number of technological forms – robots, cyborgs, androids, augmented animals, eugenically or genetically enhanced superhumans.

The modern version tends to be made of software – usually not programmed directly, but instead generated by an iterative process of selection in a carefully-designed virtual environment. Its designers create it, not though instantaneous divine fiat, but through the same process of evolution which we believe gave rise to ourselves.

Soul Revolution

Some philosophers argue that these artificial thinking-systems wouldn’t be truly conscious – their behaviour would merely be the result of billions of deterministic physical interactions at a microscopic level. Unfortunately, the same argument applies to human consciousness.

If we believe, as many of us do, that human beings are self-aware, capable of free will, responsible for their actions, and in possession of an immortal soul… there seems no reason why the same wouldn’t be true of an AI.

Even if human souls are ‘granted’ by the grace of God, it would take a very confident (or closed-minded) theologian to state that God hadn’t extended the same generosity to humanity’s creations as to God’s own.

Indeed, this distinction itself might be a false one. If human beings are created in God’s image, then wouldn’t our attempts to reproduce our own intelligence also bear that image? Wouldn’t these second-generation copies ultimately be God’s creatures as well?

My instinct has always been that AIs of whatever kind – self-aware moral agents like ourselves, brought into being by our own conscious choices – would be our children, no less deserving of love and care than the biological kind.

The next interesting question is how this applies when we recreate ourselves as software – upload our minds into virtual storage, to interact with our AIs or act as backups in the event of our death.

But that, as ever, is a question for the future.

• Read Philip Purser-Hallard's blog