Can Machines Be Creative? - University of Toronto Magazine
University of Toronto Magazine University of Toronto Magazine

Can Machines Be Creative?

The question is challenging both computer scientists and psychologists Read More

Computers can do all sorts of things once considered the exclusive domain of humans. But will they ever be truly “creative”? The question is challenging both computer scientists trying to build intelligent machines and psychologists trying to understand the workings of the human mind.

Recent years have brought what appear to be breakthroughs in artificial creativity. A Princeton psychologist has programmed a computer to improvise modern jazz, while a professor emeritus of the University of California, San Diego, has created a machine that produces works of art. But is this real creativity – or are these machines simply carrying out orders fed to them by their human programmers?

John Vervaeke, the associate director of U of T’s program in cognitive science, says the answer isn’t as clear cut as one might expect.

In some ways, the human brain and a computer are similar, he says. Both handle “symbolic processing,” such as adding up a list of numbers or reading a sentence. Similarly, humans and computers are both capable of “distributed parallel processing” involving tasks with many interacting variables, such as predicting the weather. But there’s much more to human cognition than these reasonably well-understood processes.

“There are aspects of mentality that I don’t have a clue how to program into a computer,” Vervaeke says. “Consciousness: nobody knows what consciousness is. Intentionality: you have mental states that are about things, you think about things. How that works, we don’t know.” Another is the problem of subjectivity, the fact that every human being has a unique view of the world. “The degree to which consciousness and intentionality and subjectivity play a role in creativity – and I imagine they’re huge – I have no clue, and nobody else does either.”

For now, tasks such as speech recognition or programming a robot to navigate a complex environment are keeping researchers busy. Today’s computers are fairly good at solving clearly defined, quantifiable problems, but struggle whenever a problem is ill-defined – the kind where true creativity would come in handy.

Most Popular

Leave a Reply

Your email address will not be published. Required fields are marked *