Australia’s legal system ‘not ready’ for AI as report warns of human rights concerns
Australia’s legal system is “not ready” for AI as new generative models pose risks to privacy and political communications, a leading human rights organization has warned.
Popular artificial intelligence systems such as ChatGPT are “incompatible with the right to privacy”, Amnesty International warned in a scathing worldwide report published on Thursday.
The human rights organization also claimed that AI models pose risks to “freedom of thought” by compromising broader rights concerns, including discrimination against historically excluded groups and the ability to “manipulate user intentions”.
The report called for urgent action, including from Australia, against the “most harmful practices” of AI companies that are being allowed to “produce high-risk products” at the expense of users and human rights.
Andrew Witheford, International and Crisis Leader at Amnesty International Australia, said Australia, lacking its long-awaited Human Rights Acts, was exposed to artificial intelligence.
“We have real problems because we basically have a few minor protections in the constitution,” including the single implied right of political communication, he said.
“Fundamentally, our legal system is not ready for AI. It is not ready to provide these protections.”
Mr Witheford said the Albanian government had shown a “soft touch” when it came to artificial intelligence.
“There is a strong argument for more government regulation in this area,” he said. “The government has decided that we will rely on the goodwill of tech companies.
Mr. Witheford said much of the debate about AI focused on economic impact, but the models also affected “society’s access to information, the right to privacy, what people see and hear, and public debate.”
“The potential for exploitation is very high,” he said.
“We all have the right to think what we want, but let’s not kid ourselves; it’s based on what we read, hear and see, and this will change that.
He continued: “Ultimately, Australia faces the same AI challenge as every other economy and society, but this will be exacerbated by the government opting for a light touch – essentially no touch – regulatory model.
“This presents significant dangers because it will change the way our societies work, the way our economies work, and that will affect the way our democracy works.”
Reflecting on the rise of social media, Mr Witheford said using a lighter touch was “very dangerous in the beginning”.
“You basically end up with systems that are deeply entrenched in your economy and your society… while it’s not impossible to fix, it’s harder to fix,” he said.
“We have a problem with social media. That’s why we decided to protect children. But will there be a problem with adults too?”
Mr Witheford said biases introduced into AI would “amplify existing inequalities”.
The report also raised concerns about the environmental impacts of artificial intelligence.
Asked about the impact of AI on Wednesday, Employment Minister Amanda Rishworth said it was “everyone’s responsibility”.
“Employers, governments and organizations are working to make sure we’re upskilling our workers and making sure they have the skills that will enhance their jobs… What we need to do is make a national effort to make sure we’re upskilling workers so that we can take advantage of AI, the productivity advantage.”
The Albanian government launched its AI action plan late last year, which aims to focus on the possible benefits of AI while ensuring adequate protection.
In April, Microsoft announced a $25 billion investment to power Australia’s AI-powered economy and train three million Australian workers.
