Can AI be creative?

Given that there are many humans who would not pass a Turing test, whether or not it’s the ideal yardstick for artificial intelligence is surely debatable.

Certainly as of now, contrary to Turing’s own predictions that by 2000, AI would exist that could fool enough questioners into thinking it was human, there aren’t any examples.

“No one has yet created an artificial intelligence,” says Michael O’Neill, ICON full professor of business analytics in the UCD School of Business and a founding director of the UCD Natural Computing Research and Applications Group.

Whether or not technology can be creative, a hallmark of human intelligence, is a question his teams have been working on for a long time.

What intelligence actually is, both in relation to humans and other living organisms, is a deeply philosophical question, he says.

In terms of creativity, which is a property of intelligence, there are additional barriers. “It’s not just questions of what it is to be creative but the question of whether, if a machine were to become creative, would we even recognise or appreciate it?” he asks.

With art and music, it is our fellow humans that interpret something as creative or not, he points out. “Creativity doesn’t happen in a vacuum, it has to be interpreted as such.”

He says the ultimate goal is to create artificial intelligence “but while [scientists] haven’t succeeded, along the way we have created tools that humans can use,” he says.

These include tools that can help humans to be creative, such as architects using computer aided design (CAD) software.

Innovation is also creative and here too technology can help, such as in the ability to parse monumental volumes of data to identify patterns that humans simply can’t see.

NCRA researchers won the 2019 HUMIES Gold Award for their work demonstrating human competitive artificial intelligence using evolutionary computation.

While no one as yet knows what “the recipe is for AI”, O’Neill says, it’s the direction the world is headed in. This is a fact backed up by the establishment last year of the EU’s High Level Expert Group on AI, which recently published ethics guidelines for trustworthy AI.

“AI is full of big questions around ethics and social responsibility, and these conversations are happening,” he says.

This has ramifications for consumers too.

‘Tsunami wave of technologies’

“Right now, there is a tsunami wave of technologies for business breaking on our shores, including AI and big data. Yet as customers, our experience with it is pretty dismal,” says Maurice Mulvenna, professor of computer science in the School of Computing at Ulster University.

“We are being permanently targeted by marketing communications and yet when we ring up to change something we buy, our customer experience is dismally bad, and it’s hard to find a human being to talk to. We’re fed up ringing call centres and being told our call is important to them. It’s ‘duplex’ communications and it’s difficult for consumers.”

In such situations, all we want is a conversation. “That is, I say something, you say something, and together we achieve a consensus. Trying to do that with a call centre is very difficult.”

Ironically, the advent of tech solutions such as webchat can help, because “instead of holding a phone to our ear for 20 minutes we can go and come back. It allows us to multi task.”

Customer satisfaction rises while contact centre staff are freed up to do more complex tasks. Ulster University researchers are currently working with B2B marketing specialist Maverick, in Ballina, Co Tipperary, on the application of chatbot technology to improve customer experience. Other benefits to such tools include cost savings and productivity gains.

“What we are seeing is conversations supported by AI, companies use a chat box library of good responses to customer questions, for example. It accesses its database and recommends an answer to give,” says Mulvenna. “That’s a creative solution that supports the skilled human.”