AI without conscience: a clear and present danger

A robot that achieves human consciousness: it’s a classic science fiction theme, inviting philosophical contemplation.

Are humans foolishly generating unnatural offspring? Will our digital children turn against us?

What will the human race look like from the standpoint of an artificial intelligence that becomes conscious? Will it want to destroy us?

These are just some of the questions that have fuelled the Terminator film franchise. It also adds time travel into the mix, bringing the robots of the future into today’s world.

In the latest instalment, Terminator: Dark Fate, we also witness a remarkable dramatization of a philosophical debate about the possibility of machine consciousness.

Philosophers have used the term “weak AI” to describe an artificial intelligence that follows a program of instructions and yet does not really understand what it is doing.

“Strong AI” would then be a machine that actually achieves human consciousness, able to think and feel with a self-understanding like our own.

There is perhaps no better example of weak AI in science fiction than a Terminator robot from the future who is relentlessly trying to terminate human life.

Following the instructions of its computer program, the Terminator will not stop until it has killed the human life that it has been programmed to kill.

But the tale is not simply an imaginary horror story, because today there are governments and corporations working on building military weapons that use AI.

Madam Secretary, the best television show for contemplating contemporary political themes, recently dramatized AI as an urgent problem.

In Episode 3 of the currently airing Season 6, called Killer Robots, the ethics of autonomous unmanned weapons is highlighted as something that no longer belongs to the realm of science fiction.

The best science fiction, however, has always been preparing us for this day, when the threat becomes real. Science fiction is a vivid playground of philosophical thought experiments, testing the limits of ideas in metaphysics and ethics.

In Terminator: Dark Fate, a new Terminator robot, called the Rev-9, is the android assassin from the future: an autonomous unmanned weapon programmed with a mission to kill.

The Rev-9 is thus weak AI gone wrong. The killer AI program is out of control.

But can we use technology itself to counter the threat of technology? In Terminator: Dark Fate, we are invited to entertain the possibility of strong AI in the person of a T-800 Terminator robot named Carl.

The movie holds out the possibility that Carl, who was once a programmed assassin with weak AI, has now grown a conscience. In other words, he has now truly become an android with strong AI.

While he admits that he cannot fully feel love in exactly the same way that a human being feels love, he is shown to be making the attempt.

When the Rev-9 and Carl the T-800 do battle with one another, what we have dramatized for us is really the contemporary philosophical battle over weak AI and strong AI.

Can we create a strong AI like Carl that achieves an approximation of real understanding, who thereby becomes our ethical ally? Or is it only possible to create weak AI like the Rev-9, which will inevitably be a real threat to human life?

As Madam Secretary dramatized in Killer Robots, when humans abandon ethical considerations, the threat of weak AI is a clear and present contemporary danger.

Some may hope for a science fiction happy ending, in which an android saviour with human consciousness and a heart of gold helps to save the day. If so, they will be disappointed when they learn of the “Chinese room” thought experiment of the philosopher John Searle.

Searle asks us to imagine a human who understands no Chinese and is locked in a room. Questions written in Chinese are sent into the room through a mail slot, and the human writes down answers in Chinese and sends them back out through the mail slot.

But remember the human understands nothing of Chinese. The human doesn’t even know that questions are being submitted and answers are being returned.

The human simply follows a rule book of instructions, mindlessly following a script for manipulating symbols, and fools the people outside the room into thinking that somebody inside really understands Chinese.

No weak AI can ever become a strong AI, suggests Searle, just like the human in the Chinese room who only follows a book of instructions will never understand Chinese. Our hope for the future therefore cannot be strong AI.

The character in Terminator: Dark Fate named Grace, however, suggests that grace can build on nature (as St. Thomas Aquinas would say): technology can benefit humanity, but only if we make it a part of ourselves by subordinating it to truly human ethical considerations.