MATTERS OF PUBLIC IMPORTANCE
Mr WILKIE (Clark) (15:49): I thank the member for Curtin for bringing this matter of public importance to the House. Let me start by saying that I don't see AI as either inherently good or inherently bad. However, I do believe that with the right governance, frameworks, and regulation in place we can ensure that AI enhances our lives.
The problem is, though, that for too long we've buried our heads in the sand and allowed the technology to be driven by those seeking power and profit—although it's now precisely these same people warning us that there is a 'greater than 10 per cent chance' that AI will kill all humans and that a 'swarm' could be capable of taking over the entire internet with a persistent botnet.
No wonder many people feel that this technology and the infrastructure that supports it are moving too fast and that too many decisions are being made without community consent. Indeed, I hear constituent concerns about privacy, data security, the environmental footprint of data centres and what these facilities mean for the communities that host them. There are reports of AI systems producing unexpected outcomes and concerns about the use of artists' work, writers' content and personal images to train AI models without consent—and this is before we even confront the darker side of the technology, like frauds and scams, mis- and disinformation and AI-facilitated abuse and violence.
Frankly, it's also not lost on some people that the internet has its origin in research by the US military to create a communications network that could survive a nuclear war. To shut it down you'd need to turn off every power source globally and cut every wire, microwave, fibre optic, satellite and radio link, also globally. In other words, the internet is virtually indestructible, even though it's how AI communicates and spreads and is something we might need to be able to control much more effectively if we are to prevent that rogue AI communication and spread.
All of this is quite alarming, and we would be foolish to ignore it, despite Donald Trump's unhinged comment that all the warnings are a 'hoax'. However, calls to stop development or shut everything down would seem to be the epitome of shutting the stable door after the horse has bolted. In any case, we need to remember that AI can also enhance our lives.
For instance, United Nations reports show that AI is contributing to medical breakthroughs, including speeding up drug discovery and vaccine development. AI has also facilitated early detection of breast cancer, is improving lives by making technology more accessible for people with disability and is broadening opportunities for education and mental health support.
Moreover, the boundaries of AI are complex, and it's no longer simply a tool, increasingly working its way into every aspect of our existence. In 2023 an AI-generated picture won the open creative category at the Sony World Photography Awards, and last month an individual won a landmark Fair Work Commission ruling by using an AI agent in lieu of a legal team.
Mind you, it's also the case that a survey found that one in five young people had ended a relationship in favour of an AI chatbot! Another aspect of all this is that AI transcends international borders, and the reality is that there are powerful commercial and geopolitical incentives driving this AI arms race. While America and China hold the lion's share of AI computing power, the impacts are global and should be addressed as such.
If leading AI scientists are warning that advanced AI could pose risks on the scale of pandemics and nuclear war, surely the answer is to take a precautionary approach, not to race ahead without limits. When the world confronted the threat of nuclear war, nations developed international rules and safeguards to reduce the risk of catastrophe, and that's what's needed now.
AI presents a similarly complex transnational challenge, making it clear that robust regulatory frameworks, both nationally and internationally, are needed and that they should be grounded in precautionary principles. The world cannot afford to remain on the sidelines while China and the US compete to develop increasingly powerful AI systems. We need international guardrails governing where and for what purpose AI is deployed, before the technology outpaces our ability to manage its risks.
We must demand and invest in safety by design, not simply regulation at deployment. And we must look to global collaboration to safeguard the pace of development. AI's potential is undeniable.
Ensuring that it enhances society rather than undermines or destroys it is one of the great political challenges of our time.