The artificial intelligence ‘revolution’ ensures our economy evolves

The risk to Australia from artificial intelligence is not that it will ­destroy jobs but that we might fail to take advantage of its benefits, says Paul Dalby.

Complexica managing director Matthew Michalewicz. Picture: Dylan Coker.

Nearly three decades ago, when now ubiquitous accounting software Excel was launched, there were grave fears it would destroy jobs and wipe out the livelihoods of tens of thousands of bookkeepers whose trade it would replace.

In a situation analogous to the one we now face with the recent launch of OpenAI’s ChatGPT (chat generative pre-trained transformer) and various other artificial intelligence and machine-learning-based technologies, Excel was set to automate a whole cohort of jobs out of existence.

And it did. But what happened next, and what tends to happen each time a wave of automation and efficiency sweeps through an industry, or indeed the economy, should be the lesson we take away, inspiring us to lean into this new technological revolution, experts say.

Bright Machines’ Silicon Valley-based chief executive, Amar Hanspal, in a 2021 article published by the World Economic Forum about automation, says the number of accountants working in the US was about 299,000 in 1979, six years before the release of Microsoft Excel. By 1989, despite fears of the industry being decimated, that had grown to 524,000, and about 1.28 million by 2021.

Perhaps more surprisingly, the arrival of automatic teller machines followed a similar trajectory, Mr Hanspal said. The number of bank tellers doubled from 1970-2019, despite 400,000 ATMs being in place nationwide.

“What happened? Basically, both accountants and tellers started doing more complex work,’’ he wrote. “Accountants started to spend less time number crunching and more time building client relationships, then walking those clients through more complicated problems.’’

Paul Dalby of the Australian Institute of Machine Learning.

It’s an analysis backed by Paul Dalby, from the Adelaide-based Australian Institute for Machine Learning, who says the risk to Australia from AI is not that it will ­destroy jobs, but that we could fail to embrace it and take advantage of the benefits it brings.

“Excel created jobs and it grew the industry, but the people who didn’t use Excel lost their jobs,’’ Dr Dalby said. “It’s the same with AI. Automation almost always results in more employment, rather than less, because it makes people more productive, and that part of the economy grows because it’s more valuable.

“But if you don’t get trained in those automation tools and productivity-enhancing tools you’ll lose your job to someone who has them.

“So the job for Australia is to do this rapid upskill in people being able to use all of these skills, and we’re right at the start of the revolution. It’s like being back at the start of the 1970s with computers going ‘I wonder what this can do’.’’

Adelaide-based AI company Complexica for the past few years has been showing exactly what these powerful new technologies can do with the rollout of products such as Larry the Digital Analyst and Decision Cloud to customers such as Dan Murphy’s owner Endeavour Group, Dulux Group, Pfizer and global wine company Pernod Ricard.

For a beverages sales rep, for example, the software will provide insights at the start of each work day about which products to sell to which clients, based on an analysis of data such as past buying patterns, social media analysis, time of year and demographics.

These insights even stretch as far as potential conversation starters for sales people.

Complexica managing director Matthew Michalewicz said engines such as ChatGPT were a natural evolution of the search function.

Three decades ago, he said, digging out complex information likely involved heading to your university library and trawling through the stacks until you found what you needed – a process taking many hours, if not a day or two.

Search engines, most notably Google, up-ended this paradigm, and ChatGPT is the next step in this process.

“With Google there’s this massive acceleration of our ability to access knowledge, and when you look at what OpenAI has done it’s kind of the next level of that,’’ Mr Michalewicz said.

The next level up from what Google provides – a series of links – is “conversational answers which are much more in-depth and contextual’’, such as those that ChatGPT now offers, he said.

ChatGPT is a conversational artificial intelligence software application developed by OpenAI. Picture AFP

“Not just links to web pages but something that has analysed and drawn on several sources of information and delivered that back to you,’’ Mr Michalewicz said.

“I think this trend is irreversible and goes in one direction.

“When you look at this type of algorithmic technology, it will be as pervasive, entrenched and ingrained absolutely everywhere, whether you’re an engineer, a pilot, a sales rep.

“I see what we are doing, what OpenAI has done, the Google search engine, all of these kind of things, as kind of one continuum … augmenting human capability in access to information, access to knowledge, decision-making.’’

Mr Michalewicz said it was natural that schools and universities were worried, as the process by which we gather information, generate and test hypotheses and produce outcomes such as a university theses could now be automated to a large extent.

Much as many of us have already lost the ability to do simple arithmetic in our heads, these skills could wilt; however, the lesson from history was that humans move up the value chain in terms of how they spend their time.

“No one wants to go back to doing more analysis, we like the comfort of this automation,” Mr Michalewicz said. He said it was also useful to keep in mind that AI algorithms were still quite “dumb” in some ways, and there were many things they could not do.

An example of this was that image recognition algorithms still had difficulty differentiating between a picture of a chair, or a picture of a reflection of a chair – a distinction immediately apparent to a human observer. But algorithms are particularly well suited to processes involving large data sets, or for automating repetitive tasks.

One of the huge benefits already flowing into the economy was the opportunity for small businesses to use algorithmic tools to automate several previously time-consuming processes, Mr Michalewicz said. An example of this was the already affordable and common use of chatbots on small business websites to respond to inbound queries.

“It makes starting a business or operating things at the smaller end of town much more accessible, much more affordable, and that creates jobs,’’ Mr Michalewicz said. He said jobs that were repetitive and involved the collection of basic knowledge would be those most at risk of replacement by AI.

“Making an inquiry with Qantas, for example; they are very simplistic in terms of the queries so that can easily be automated.’’

Complexica in the past 18 months has raised $25m from investors including Perennial Partners and Microequities Asset Management, with the money being committed to rapidly growing the business.

Matthew Michalewicz’s company Complexica automates wine production. Picture: Russell Millard

Other uses of AI operating in the field include Adelaide company Life Whisperer’s technology that helps fertility specialists select the most viable embryos, and LBT Innovations’ high-speed classification technology for culture plates used in laboratories.

Dr Dalby said currently we don’t have much clarity as to what extent AI and machine learning will be embedded in each industry; however, it was inevitable it would touch most areas of the economy. Those business and sectors that don’t embrace it will be left behind.

“Australia didn’t invest in automating in the car industry like other countries did and that didn’t save us jobs, that cost us the industry,” he said. “It’s a bit of race to see who are the adopters of this technology. Those businesses, sectors, workers are going to thrive and those who don’t are going to fall by the wayside, so it’s a bit counterintuitive.

“You have to adopt the automation to keep the employment growing.’’

Dr Dalby said there were “literally thousands of ways’’ the technology could be applied, and his own institute was a fertile recruiting target for sectors such as the creative arts.

Automating parts of the movie production process, for example, “doesn’t mean that there are less jobs, it just means there are more movies with cool stuff’’.

He said while there had been some controversy around the use of AI in generating art, this reflected a similar outcry when devices capable of generating electronic music hit the scene decades ago.

And in that case, an entire new genre of music was born.

But humans would remain integral to every business, Dr Dalby said. “It’s very hard to see it replacing humans. In the end, a human has to do the final approval, the human has to interact with customers, it’s still going to be people in the system, but they’re doing higher-value things and they’re probably going to get paid more.’’

PwC’s Sizing the Prize report, which looks at the impact of AI globally, estimates a huge potential contribution to the economy of $US15.7 trillion and up to a 26 per cent boost in GDP for local economies by 2030.

Originally published as The artificial intelligence ‘revolution’ ensures our economy evolves