Artificial intelligence has prompted a supercharged tech boom, with valuations of the world’s biggest companies soaring on the Nasdaq amid talk about a new era of human-like algorithms.
Here in Australia, a microcosm of Silicon Valley is happening, with 30,000 people now working in AI and many start-ups tapping into billions of dollars of venture capital.
One local start-up that just got $2.5 million in seed funding is Apate.
The company, originally founded out of Macquarie University, has clients including the Commonwealth Bank that pay Apate to unleash chaos onto global scammers with tens of thousands of AI-powered chatbots.
Dali Kaafar’s AI company has tens of thousands of chatbots with different “personalities”. (ABC News: John Gunn)
Founder Dali Kaafar came up with this idea several years ago, after he picked up a cold phone call from a scammer and decided to keep them on the line by playing “naive” and “confused” for 44 minutes.
“The scammer’s time was completely wasted. A very unproductive phone call for them,” he laughs.
Out of the university, Professor Kaafar developed a large language model (LLM), which was trained to chat with scammers using videos taken off the internet of real-life conversations between scammers and dupe actors.
“Pretty much to learn how to really react and to engage scammers,” Professor Kaafar explains.
Mr Kaafar has developed a large language model, which it trained to chat with scammers. (ABC News: John Gunn)
Today, Apate has tens of thousands of chatbots with different “personalities” and even languages. It deploys them by posting phone numbers on the internet and dark web that scammers unsuspectingly call.
“They have different attitudes, different personalities,” Professor Kaafar says.
“Some of the conversations can be pretty much shocking to some extent, if they’re rude or sometimes they’re very, very polite. Sometimes they’re way too cooperative.
“They’re scamming the scammers themselves.”What other AI is being developed in Australia?
Both here and globally, as AI advances quickly, even defining what it is is becoming more complex.
Rob Nicholls, who is a senior research associate at the University of Sydney’s Centre for AI, Trust and Governance, sees the AI currently being talked about as a continuum of decades of software development.
“Strictly, artificial intelligence is essentially a way of getting a computer to do things which are more complex than what traditional software did,” he says.
“I see it as being on a continuum, but it’s had a step change.”
Like many in this space, Dr Nicholls views the release of ChatGPT by US company OpenAI in November 2022 as the moment that supercharged AI and brought it into homes and everyday conversation.
“Actually releasing a chatbot version of a large language model changed the perception,” he argues.
“What was actually available to computer scientists and people who played with artificial intelligence in academia and in business was suddenly available to everybody.”
AI start-up Heidi Health has been developing its platform for the healthcare industry. (Supplied)
Over in Melbourne, another AI start-up Heidi Health has been developing its platform for the healthcare industry since before the ChatGPT explosion.
The AI-powered app lets doctors and vets record conversations with their patients (or their owners) and then it generates notes, such as referrals to other clinicians, surgery notes or patient history summaries.
“It’s kind of listening in the background,” Heidi Health’s clinical director Ben Condon explains to ABC News.
“It’s not actually diagnosing.”
Less than four years after it was founded, Heidi Health now has major clients globally, including hospitals here in Australia and the owner of Petbarn, Greencross. It also just got a capital injection from the fund Blackbird.
Ben Condon says he believes this AI platform can save doctors’ time. (Supplied)
Melbourne-based Dr Condon came over to the company this year, after giving up his full-time job as a doctor on emergency wards in hospitals. He says he believes in the company’s mission to save doctors time.
“I’m really excited by the chance of Heidi to make work a little bit more sustainable and hopefully keep clinicians seeing patients and prevent them from burning out.
“It’s been a wild ride, but it’s been lots of fun.”AI is becoming entrenched in tech
The Australian tech industry’s lobby group says all of its 180 members, which include big names like Google and Atlassian, are today using AI in some way.
The Tech Council of Australia is pushing for further investment, including from Australia’s massive superannuation funds, and has a goal of growing AI jobs exponentially to 200,000 by 2030.
Damian Kassabgi says he sees AI as a big opportunity. (ABC News: Geoff Kemp)
“There are around 30,000 people whose main job is they wake up every day and they work on AI,” TCA chief executive Damian Kassabgi tells ABC News.
“We believe that grows to 200,000 by 2035, if we get the investment right.”
The opportunity isn’t just in AI start-ups, he says, but also in supporting jobs in other related areas, like the data centres being erected to store all of the information needed for AI systems.
“From our perspective, there’s a huge opportunity here,” Mr Kassabgi says.
“And the question for Australia is what is our piece of the pie here?”
“We are just starting this next stage of the AI revolution and that’s being led by not just Nvidia, Microsoft, Palantir and others,” says US-based analyst Dan Ives, from Wedbush.
“We don’t believe this is a bubble. This is the fourth industrial revolution.”
However, some are raising concerns about hype.
In the US, for instance, the privately held OpenAI is reportedly in talks that would value it at $US500 billion — bigger than Elon Musk’s SpaceX — but it has never actually turned a profit.
Other companies are in renaissance off the back of AI, such as the 48-year old US company Oracle, which recently surged in value to briefly make its boss Larry Ellison the world’s richest person, surpassing Elon Musk.
Some concerns have been raised about the momentum of AI. (Getty Images: Cesc Maymo)
“A rational investor must short this,” quipped JP Morgan’s former head strategist recently about Oracle’s surge.
There are also concerns that AI isn’t as revolutionary as its creators claim, with one viral report by the US university MIT recently concluding that many companies are struggling to use it in productive ways.
The University of Sydney’s Dr Nicholls believes there is hype around AI, and that it is being “amplified by algorithmic trading” and “the temptation” to have good news about AI stocks.
“A sort of essential part of that hype is that we’re on the way to artificial general intelligence now,” he says.
Dr Nicholls still believes the industry is 20 or 30 years off truly mimicking humans, and that AI as it currently stands is immense data systems that are very good at predicting trends and patterns.
“What it doesn’t do is have creative thought, and that’s a really tough thing to teach a machine to do,” he notes.
AI is ‘coming at us fast’
Another academic who works in the AI space, Belinda Barnet from Swinburne University, shares concerns about an over-hyped sector, and worries that a “cultish” industry is going to power ahead without “guardrails”.
“They’re selling themselves as the natural future and that we don’t have much choice. It’s coming at us and it’s coming at us fast,” she says.
“We need to ask some serious questions about what the social effects of this technology is going to be.”
One of the big discussions in the AI space in Australia right now is whether companies should have to pay to access all of the information that they’re training their systems on, including works by artists and writers.
“If these companies, for example OpenAI, had to actually pay creators and pay artists and pay writers for every piece of information that they are training on, they’d go broke really quickly,” Dr Barnet says.
“They want everything for free.”
Authors warn against AI copyright exception
The Tech Council’s Mr Kassabgi says his members hope the AI industry finds a “middle ground” with concerned artists. Generally, he says the conversation around regulation is “not necessarily about changing individual laws”.
“This is about ensuring that regulators and government understand AI and actually understand how to implement the laws in a new world where there’s AI,” he says.
Apate trained its chatbots using videos hosted on YouTube and Vimeo. Professor Kaafar says the company got the permission of those platforms to use the content but not the original creators.
Professor Kaafar says Apate would be open to a model that compensates creators whose work is used to train AI. Generally, he also hopes his company will become profitable within years.
“We’re really building what we like to call a counterintelligence platform,” he says.
“We’re weaponising AI for good.”
This is the first part of a week-long deep dive into AI airing on ABC TV’s The Business. You can watch this story on tonight’s episode at 8:44pm AEST on ABC News and 10:30pm on ABC, or any time on iView.