Julian Hill (Bruce, Australian Labor Party) Share this | Hansard source

I move:

That this House:

(1) notes that the development and implementation of artificial intelligence (AI) is rapidly accelerating globally and in Australia;

(2) acknowledges that while there is much uncertainty surrounding both the development and adoption of AI technologies, and that 'AI' is a term used to describe a variety of techniques and applications, what is clear is that these technologies will transform human society, how we experience our lives and how we understand reality;

(3) recognises that harnessing the benefits of AI presents enormous opportunities for Australia, including:

(a) the potential for AI to boost productivity and revolutionise many industries;

(b) the capacity to transform our economy with advances in every conceivable field of human endeavour;

(c) new employment opportunities through human-centered AI;

(d) improving health, wealth and equality outcomes for all Australians including through improved government service delivery; and

(e) enhancements to environmental sustainability through better-informed decision making and accelerated scientific discovery;

(4) further notes that in order to safely harness these benefits, Australia must also act to mitigate the profound risks posed by AI, including:

(a) immediate and tangible threats to job security and industrial relations;

(b) the risk that AI could perpetuate or amplify existing biases and discrimination;

(c) the risk that AI could perpetuate or enable new forms of disinformation and misinformation;

(d) social and democratic harm through the use of AI in cyber attacks and large-scale disinformation campaigns;

(e) further digital marginalisation and inequality; and

(f) the threat of social disruption and national security risks;

(5) recognises that notwithstanding positive efforts underway to address matters related to AI—including responsible AI standards and policy—Australia has broader capability and governance gaps and needs to ensure that regulatory oversight of AI development and adoption in Australia is fit for purpose;

(6) affirms that:

(a) AI is one of the most transformational technologies of the 21st century, on par with the industrial revolution;

(b) the level of risk posed by unchecked AI, and the scope of policy development needed to curtail this risk, warrants urgent attention;

(c) industry leaders are calling for additional government action and regulatory cooperation;

(d) AI governance, regulation and public-good investment is too important to be left to industry or technical experts alone; and

(e) the Australian Parliament and Government have a responsibility to consider and act thoughtfully and promptly in responding to these changes; and

(7) further notes the recent regulatory moves underway in other jurisdictions, including diverse approaches to AI governance in the EU, the USA, China and the UK; and

(8) recognises that all Members of Parliament have a responsibility to engage with the transformative challenges presented by AI, and together explore what Australia should do to:

(a) foster and contribute to a national debate about AI;

(b) seize the enormous opportunities that AI technology will continue to generate;

(c) mitigate, through appropriate regulatory measures, community anxieties and the profound risks posed by unchecked AI; and

(d) deliver an Australian approach to AI governance and regulation informed by values of democratic participation, nation building, social justice, equality, consumer protection and international cooperation.

It's so important that MPs engage and debate this. ChatGPT has fuelled public awareness of artificial intelligence, but large language models are just the canary in the coalmine. AI technologies are set to transform human society, how we experience our lives and how we understand reality. Like a new pair of glasses or hearing aids that are difficult to remove, AI will shape our perception of life as it influences what we see, hear, think and experience online and offline. Our everyday life will be augmented by having a superbright intern always by our side. Service delivery by governments and businesses will be transformed, unleashing creative destruction in some sectors of the economy but hopefully driving Australia's next big productivity boost.

Our understanding of the universe and the world has been developed for millennia through reason, the scientific method and faith. Yet over the next generation, living with non-human pseudo-intelligence will challenge established notions of what it is to be human. Advanced AI systems trained on massive datasets are already spotting new links and patterns. Many appear true but are beyond human comprehension. Others are downright wrong, as a result of faulty data or algorithms. AI is maths. It's not magic. It's not a modern-day Delphic oracle to be worshipped or blindly obeyed. Citizens and policymakers must urgently get a grip.

While AI has the potential to generate enormous wealth, exponentially more powerful AI technologies unaligned with human ethics and goals bring unacceptable risks. Put simply, superpowerful AI is a sociopath, a shoggoth disguised with a smiley mask by engineers trying to align it to prevent misuse. But how might authoritarian governments, bad actors or rogue citizens misuse AI to trash human rights, attack democratic societies, scale up scams and organised crime, and harness dangerous knowledge and capabilities otherwise unavailable? It's more serious than the opposition's big fear, of course—that ChatGPT may support the Voice to Parliament! But, just as humanity was right to worry about the risky gain-of-function virology research, we're right to worry about uncontrolled generative AI. Imagine unleashing this pseudo-intelligence with self-executing power connected to the internet without intermediating human judgement.

Safer AI technologies, well aligned to human values and needs, are only possible with public intervention. Decisions that shape the future of our society cannot be left to the private interests of technologists and multinationals alone. Governments must act in the public and the national interest to establish guard rails and determine how and where to apply both the accelerator and the brake, harnessing the benefits of AI while mitigating and managing the risks.

AI use must support our Australian, established values of social justice, equality, democratic participation and nation-building. Our new government is now acting to address Australia's AI capability and governance gap. Australia has the chance now to cherry-pick from world-leading and diverse approaches globally and to craft our own world-leading AI regulatory response, one which unashamedly voices, champions and embeds our values and the famed Australian commitment to a fair go. None of this is easy.

In saying this next bit, I point out that it's not government policy. I propose the establishment of a time-limited AI commission, just for five or six years, located right at the centre of government and bringing together industry, public servants, academia and civil society. Its functions could include fostering public awareness and education; rapidly building capability across the public sector; formulating options to guide policy responses, including sovereign capability; reviewing proposals; and engaging with the leading international thinking and debate around AI. What you might call, in a non-catchy phrase, a light-touch, cross-disciplinary institutionalisation would see ethicists, lawyers, philosophers, psychologists, economists, doctors, sociologists, educators and public administrators as equal partners with the scientists and the technologists. We can't support pushback from scientists and the multinationals saying this is too complex. Parliamentarians have to be engaged with this debate. None of it's easy. A parallel could be drawn with the Climate Change Authority, the Australian Cybersecurity Centre or the Canberra Commission on the Elimination of Nuclear Weapons.

I believe we also need to regulate access to the most powerful tech. We don't allow people to pick their kids up in battle tanks or businesses to run around with rocket launchers, and I'm pleased the government is acting now.

See this speech in context