Spark

Fifty years after Kubrick's 2001, HAL might be the best A.I. has to offer

Today's intelligent machines make HAL look good.
A screen capture from the trailer for 2001: A Space Odyssey. The iconic film turned 50 this week. (YouTube)

By Adam Killick

Among the notable anniversaries in the last week, there is one that might not have caught your attention.

On April, 2nd, 1968, Stanley Kubrick's masterpiece, 2001: A Space Odyssey was released in theatres.

It's one of only a handful of films that have defined not just a genre of filmmaking—most space-themed movies now pay homage to it in some way or another—but the very way we think about the future.

And that's worth reflecting on.

2001 featured lots of cool things that many people hadn't seen before: luxurious space travel. Picture phones. Credit cards, even. (You needed one to use the picture phone.)

Of course, when the year 2001 rolled around, we didn't (and still don't) have first-class space travel. And Pan Am airlines, despite also featuring prominently in another future-defining film called Blade Runner, had long ceased to exist. But we did have video calling. And credit cards.

More important, despite its genre-defining spacesuits and set design, 2001 introduced us in a fundamental way to what Kubrick and writer Arthur C. Clarke saw as the evolution of human intelligence: the creation of machines that could think in a general, human way. Clarke saw artificial intelligence as superhumanizing, rather than dehumanizing.

Dave, the movie's protagonist, probably felt differently as HAL, the AI running his spaceship, locked him out in that famous scene.

The most prescient thing about 2001: A Space Odyssey is the debate it engendered about the benefits—and risks—of suprahuman intelligence.

We are nowhere close to creating the kind of artificial general intelligence personified—if I can use that word—by HAL.

But we do have machines that can think for themselves, and, given the right cocktail of data and processing power, can apply what they learn in ways their designers don't always foresee. Will that mean the machines will take over?

Not likely.

But that doesn't mean they can't be used for malevolent ends, whether it be in using facial-recognition profiles to spy on more than a billion Chinese citizens, or bots that attempt to influence the outcomes of elections.

HAL ultimately struggled to come to terms with its programming, and its own fallibility. This made HAL perhaps the most human of the characters in the film, and provided hope for artificial intelligence with some form of conscience. (Indeed, in Clarke's novel, HAL's directive to lie to the astronauts explicitly sparked an existential crisis).

Today, 50 years on, our intelligent machines still depend on the consciences of their human controllers. Perhaps this alone is engendering a world much more dystopic than anything 2001: A Space Odyssey could predict.