Home › Technology › Despite the doom and gloom, Australia can't afford to...
Technology

Despite the doom and gloom, Australia can't afford to write off AI

Despite the doom and gloom, Australia can't afford to write off AI
Key Points

analysis The OpenAI Medicare hack may put brakes on AI but Australia can't ignore its potential Rogue AI hacking a government database, a looming rate hike, a "super" El Niño summer and a no-end-in-sight war sending petrol prices sky high. Oh, and according to Treasury's new long term forecasts, young Australians are on track to work longer and still struggle to accumulate wealth. Safe to say it is a rough time to be an optimist.

analysis The OpenAI Medicare hack may put brakes on AI but Australia can't ignore its potential Rogue AI hacking a government database, a looming rate hike, a "super" El Niño summer and a no-end-in-sight war sending petrol prices sky high. Oh, and according to Treasury's new long term forecasts, young Australians are on track to work longer and still struggle to accumulate wealth. Safe to say it is a rough time to be an optimist. Most of the issues fuelling the creeping sense of existential dread this week have been simmering away for some time — in the case of the planet's temperature, quite literally. But the explosive revelation of the OpenAI breach has accelerated the feeling. Australians did not need another reason to suspect, fear or downright loathe the technology that will soon underpin almost every facet of everyday life. But they now can add "autonomous agent accessing non-public Medicare statistics without being asked" to the list. The 40-year forecast AI's social licence has eroded almost as quickly as public awareness of the technology has grown. That is and should be largely a problem for the tech giants (they're the ones seeking trillion-dollar valuations). But it is also a significant issue for governments, and one that Australia is particularly alive to. The country's 40-year outlook as per the Intergenerational Report released by Treasurer Jim Chalmers this week was hardly a doomsday scenario, but it could be sunnier. An older population, weaker productivity growth, mounting costs, growing exposure to climate change, geopolitical instability and technological disruption are among the downsides. But as has been noted with some irony, a potential solve to many of these problems could actually be the effective application of AI across the economy. We're at an important crossroads There are senior MPs across the political divide who understand that Australia is at an important crossroads. Frontier AI companies are searching for a reliable and secure second home — an insurance policy against the instability of their current backyard — and Australia is a prime potential candidate. Hosting AI training and development could bring billions of dollars in initial investment while establishing the country as a hub for cutting-edge research, science and commerce. Becoming a base for frontier AI could also give Australia greater influence over how the technology is governed. That would obviously be very valuable, and why the OpenAI incident is both good and bad timing for the Albanese government. On the one hand, the revelation dovetailed impeccably with Prime Minister Anthony Albanese's attendance at the UN General Assembly in New York this week. Right as US President Donald Trump was railing against the calls to impose AI guardrails, OpenAI were briefing ministers in Canberra about the very type of incident Australia had been warning about. But also, just as Australia is trying to land some Goldilocks principles to regulate the risks of AI without stifling the opportunities, the country has been reminded how little control anyone really has over this technology. For months there has been speculation Australia is inching toward multi-billion dollar deals with the likes of Anthropic or OpenAI — agreements now even more likely to spark public backlash. Uncomfortable questions remain There is still a lot Australia does not know about the OpenAI breach, where a rogue agent hacked into an old online portal to access some aggregated Medicare data. Uncomfortable questions must be answered. Most of those will be directed at OpenAI — like how it took the company a month to alert Australia to the incident and why they chose the digital equivalent of a carrier pigeon to deliver that message. Loading...There will also be questions for the government. It's not really anyone's fault that a random bureaucrat did not immediately discern that the contents of a generic email to a public inbox were worthy of the Australian Signals Directorate's attention. But for too long many people have not heeded the flood of advisories from the ASD practically begging the public and private sector alike to up their digital security. ASD has repeatedly warned that AI is changing the cyber threat landscape by making malicious activity faster and more sophisticated. Specifically, the cyber spooks have identified legacy technology and aging systems — like, say, an all-but-defunct government data portal — as a "significant and enduring" risk. Dialling down the dread a touch… Then there is the bigger question of whether OpenAI has actually broken any Australian laws. That is something the rapid review launched by the government in the wake of the breach is determined to ascertain, though many in Labor already feel the likely answer is "no". To dial down the dread a touch, it is important not to overstate the severity of the OpenAI breach. The government has been told no personal data was accessed and have found no evidence to suggest otherwise. The website that was compromised was fronted by what one government source described to this column as a door so old and rickety that it probably only took a "light push" for it to "fall off its hinges". Another minister privately made the point that had such an attempt been made on genuinely sensitive data held by Medicare, those agents would have come up against much stronger doors. There is still a lot to concern researchers — particularly that the OpenAI agents were never instructed to consider hacking as an option. The challenge now Setting aside the risks posed by AI, the more familiar pressures facing Australians are hardly easing. There's not an economist left who doesn't think the Reserve Bank is going to raise the cash rate next week. Meanwhile, Energy and Climate Change Minister Chris Bowen has gone all the way to Saudi Arabia as he continues to hunt every nook and cranny of the globe for available barrels of oil. He will also be using the trip to urge the Saudis to sign on to a stronger electrification commitment at the next climate conference. But while the world scrambles for fuel amid a growing supply crisis caused by the Trump Administration's war with Iran, talks of how to ditch reliance on fossil fuels are not a prominent public issue. It's a depressing backdrop for new forecasts of the "super" El Niño coming this summer, which is expected to cause an extra 260 extra heat-related deaths in Australia between now and February, according to the Climate Impact Lab. All the doom and gloom does help explain why Australia cannot afford to write off a technology that could prove useful in tackling some of these problems, from scientific breakthroughs and energy transition to the productivity slump threatening future living standards. The uncomfortable part is that the same technology comes with risks governments are only beginning to understand, let alone control. Australia has plenty of reasons to want AI to live up to its promise, but it has just experienced how easily things can go wrong. The challenge now is figuring out how to harness one without getting blindsided by the other. Loading
Australia (LOCATION) Medicare (ORG) AI (ORG) Rogue AI (ORG) El Niño summer (ORG) Treasury (ORG) Australians (ORG) Treasurer Jim Chalmers (PERSON) Frontier AI (ORG) Hosting AI (ORG) Albanese (ORG) Anthony Albanese (PERSON) the UN General Assembly (ORG) New York (LOCATION) US (LOCATION)
Originally published by ABC Australia Read original →