Wooley: AI has its place but it’s still no match for the human touch

Artificial Intelligence may well be able to create images that on the surface look real enough. But where’s the colour, where’s the sparkle, where’s the self-awareness, asks Charles Wooley >>

This photo taken by the Mercury’s Chris Kidd of firefighters battling a blaze at a Brighton scrapyard, looks so realistic you can almost feel the heat coming off it. This picture makes others made by artificial intelligence, pale into insignificance, according to Charles Wooley. Picture: Chris Kidd

THERE’S a lot to worry about just now, but I would like to direct your thoughts away from how you are going to pay the mortgage, to the challenge of artificial intelligence.

AI, as it is known, is going to be around a lot longer than Philip Lowe, the precarious and not too popular Reserve Bank governor.

So, let’s start the diversion with a photograph released this week by a Sydney artificial intelligence art studio, Absolutely. AI.

The studio asks a timely question: “Who is the better artist, man or machine?”

For what little my Luddite opinion is worth I thought the snap they developed was lurid and unrealistic and represented no seascape I had ever seen in nature.

And what are we to make of what appears to be a giant porpoise or the legendary kraken lurking in the centre-bottom of the frame? Is it about to swallow the surfer?

But what do I know? The AI photo won a photography competition run by DigiDirect, an electronics retailer.

Cynics might suspect some (possibly unconscious) insider bias in that judgment.

This photo of waves crashing on to a beach, was created by artificial intelligence studio Absolutely.AI. And despite recently winning an amateur photography award, does not look realistic at all, according to Charles Wooley. Picture: Absolutely.AI

But the AI art studio was doing a convincing simulation of triumphant human crowing.

“We did it to prove that we’re at a turning point with artificially intelligent technology, by passing the ultimate test – could an AI-generated image not only slip by unnoticed but actually be awarded the top prize by a photography expert?

“The answer is resoundingly, yes,” the studio claimed.

Has this latest version of robotics now relegated today’s human photographers to the same fate as 20th century automotive workers, to be replaced by machines?

Not if you consider another picture published this week.

Mercury photographer Chris Kidd snapped an amazing shot of a scrapyard blaze at Brighton, for the front page of Tuesday’s Mercury.

Chris captured every element of the story in one brilliant photograph. The firefighters, the aerial appliance, the jet of water, the putrid smoke and the steam, all framed around the metallic chaos of a junkyard disaster.

My eyes were stinging, and I could even feel the heat.

In stark contrast to the photo above, this news pic taken by Mercury photographer Chris Kidd of firefighters battling a blaze at a Brighton scrapyard this week, looks so realistic you can almost feel the heat coming off it, according to Charles Wooley. Picture: Chris Kidd

Incidentally, I challenge any AI photo lab to rearrange one jot of the carnage to create a better photo than Chris managed to take by actually being there.

He should certainly win an award for that picture, so long as he can prove himself 100 per cent human.

Judges can’t be too careful these days.

Way back we had robots to worry about. For more than a century science fiction writers warned about humanoid machines rebelling against us to become our masters rather than our slaves. The word “robot” was first used in 1920 to name a mindless clanking machine that resembled a human moving inside a too-tight suit of armour, with a carrot up the bum.

Sometimes science fiction becomes science fact.

And although the machines didn’t actually become human in appearance and movement, robotic machinery did effectively displace people in a wide range of industries.

But a bigger threat was emerging with the drive to create a machine that could think like a human, only with greater computing power and with the ability to design and build even smarter machines.

Whatever the danger, science can’t help itself. It has to meddle with the world, regardless of the consequences.

A 'Photo' of Queen Elizabeth II made entirely by Sydney-based AI studio - Absolutely.AI. Pictures: Supplied

A long time ago I met a grand dame who had been at the University of Manchester in 1932 when Ernest Rutherford and his team first split the atom. She told me: “I heard them say they considered that the splitting of the nucleus might get out of control and cause a chain reaction. It had never been done before and no one knew for sure if it might destroy the whole world. But they did it anyway.”

Hiroshima, Nagasaki and whatever Putin gets up to with nuclear weapons in Ukraine, that 1932 experiment might still destroy the world.

Then again, in the age of what looks like disastrous global overheating, some experts still believe nuclear power might save the world.

We have to give science the benefit of the doubt. It could save our bacon. Or burn it to a crisp.

Just because photographer Chris Kidd scored a goal for the humans this week, easily beating the AI machine’s lurid photo-confection, we shouldn’t rest easy.

AI is writing creditable essays for university students. It writes job applications. And given that the literary bar is set so low, with almost half Australians effectively unable to read and write, the machines might be on to a winner by default.

There have been intimations that journalism will be next.

Having read this far, would you be surprised to learn that this week’s column has been written by AI?

“Better than his usual drivel,” do I hear you say?

A ‘Photo' of a small child and grandparent made entirely by Sydney-based AI studio - Absolutely.AI. Pictures: Supplied

Well, you’d be wrong. It is actually my usual drivel, and so far I don’t think any machine is sufficiently self-aware to take the piss out of itself, no matter the quality of the writing.

But having raised the spectre, let’s see what the machine-mind can write about itself.

We asked the AI service ChatGPT to write a 100-word newspaper article about itself and the dangers of AI to humanity.

Here’s what came back.

“Artificial Intelligence is a rapidly advancing field, but with this progress comes the potential for danger. As machines become smarter, they are able to make decisions and act in ways that could harm humans. There is a growing concern that AI systems could malfunction, causing physical damage or even making decisions that result in harm to people.

“Additionally, there is the risk of AI being used for malicious purposes, such as cyber-attacks or even warfare. It is essential that we develop safeguards and regulations to prevent the negative effects of AI on humanity. We must ensure that the benefits of this technology are balanced with the protection of human lives and wellbeing.”

The machine is having an each-way bet. Those are reasonable and obvious points. There’s nothing to argue with there. But there’s nothing new.

And where’s the colour, where’s the sparkle, where’s the self-awareness?

Actually, I am quite relieved.

But if 20,000 readers tell my editor that the previous paragraph was the best part of this week’s column, then I will chuck it in and go trout fishing.

There’s no machine silly enough to do that.

Charles Wooley is a journalist, writer and former reporter with current affairs TV show 60 Minutes.