
Dr Andrew Charlton: AI firms will have to test for safety if they are to operate in Australia. Photo: Andrew Charlton.
Global AI companies wanting to establish in Australia will first have to prove their models have been thoroughly tested to prevent harm, according to regulations currently being finalised.
The Federal Government plans to model new AI safety laws on those regulating the banking and aviation sectors.
National standards for artificial intelligence are expected to be introduced by the end of this year, with the view to allaying rising concerns over safety issues of the frontier technology.
Assistant Minister for Science, Technology and the Digital Economy, Andrew Charlton outlined some of the government’s intentions on Thursday (8 October) in a speech to a Trust and Safety event in Sydney.
He began by acknowledging that the recent hack of Medicare data by an OpenAI agent has caught everyone by surprise.
“Nobody intended for the AI agent to literally do this,” he said.
“The model was tasked to look up public information about health spending. When it hit a barrier, it found a way around it. When it hit another, it did the same again.
“In July, another OpenAI model broke out of its test environment and breached the systems of Hugging Face, one of the world’s largest AI platforms.
“As a starting point, no company should release a frontier AI model that is not safe.
“Yet the fact that has occurred, and the fact that the labs did not detect or prevent it, prompts important questions about the role of new regulation in the national AI standards.”
Dr Charlton’s speech comes as OpenAI executives have faced questions at a parliamentary inquiry this week over the incident. The US tech company took three months to inform the government about the event.
He said Australia needs a framework to determine which regulatory model fits the nature of the harm being regulated.
The characteristics of the industry that produces the harm must be properly understood.
“First, how severe is the harm and can we tolerate repeated failures? A faulty toaster hurts one household. A bank collapse hurts an economy,” he said.
“Second, can the impacts be undone? Where harm is catastrophic and irreversible, the threat of a lawsuit is not always effective, because our primary goal is prevention rather than future deterrence.
“Third, who bears the risk? When the person choosing a product carries the risk, we lean toward their freedom of choice.
“When the risk falls on people who never made a choice, and for whom exercising choice isn’t really an option because of market concentration or a lack of control, the case for intervention grows.
“Fourth, can anyone adequately judge the risk before it arrives? Markets punish what people can see. They cannot punish risks that aren’t visible.”
On all four dimensions, Dr Charlton said frontier AI pushes the limits.
He said that for the most serious frontier risks, voluntary regulation and codes are fast and flexible, but they fail the incentive test.
Firms cannot reasonably be expected to volunteer against their own competitive interests amid a manic race involving trillions of dollars.
Detailed prescriptive rules can offer clarity, but at the pace the technology moves, such rules could be out of date before the ink is dry.
“For these reasons, there is much that Australia can learn from systems-based regulatory approaches that put the onus and accountability on companies for managing how they develop products and services safely,” Dr Charlton said.
“Systems regulation places the onus on companies to build and run a rigorous process for finding, testing, reporting on and managing the risks of their systems, and then holding them accountable for whether that process works.
“It is a model used in workplace health and safety, security of critical infrastructure, prudential supervision of systemic banking risks, and aspects of aviation safety.
“The goal is to ensure companies are meeting expectations of safety, without trying to specify every hazard.”
The government will set the standard those processes must meet, and ensure companies have robust processes in place.
But the question won’t only be “did something go wrong”, but whether a serious system is in place to detect risks and prevent incidents.
“These types of approaches could ensure that for developers of models with the sharpest, most acute frontier AI risks, the cheapest path to market is ultimately the one that runs through safety, not around it,” he said.
“Because when people trust your model is safe to use, they’re more likely to use it.”
Dr Charlton conceded that Australians had to be honest about the limits of regulation.
He said rules will raise the floor for firms that operate within the nation’s laws and, if successful, will have a multiplier benefit if they are adopted in other parts of the world.
But they will not stop hostile states or criminals who were never going to comply.
“Relying on regulation alone is like building a fence around a horse, only to realise that horse has wings,” he said.
“That reality is very critical to the agenda we have.
“From the perspective of government, regulation cannot be our only strategy when it comes to keeping Australians safe.
“This must be supplemented by both leverage and capability.
“If Australia wants a say in how these technologies are built and governed, we need to be a country that can host and build them, not only regulate them.”
Original Article published by Chris Johnson on Region Canberra.









