Australia urgently needs to elevate its discussion of artificial intelligence from the technical to the strategic. The routine calls for more money to be invested in AI in Australia have lacked bite despite the obvious advantages Australia has in AI, particularly at the research level.
This week’s budget announcement of increased funding for AI initiatives still falls short of what industry says is required. Australia certainly lags behind many similarly sized economies for AI investment, but more depressing is the lack of vision and leadership needed to incubate a robust AI industry. At the same time we need to think about how to mitigate the serious national security challenges that AI driven digital networks pose.
The world has moved on from the national AI strategies that emerged five years ago. The trend appears to be towards developing stronger linkages between states and tech firms. Ex-Google CEO Eric Schmidt has been championing partnerships between technology firms and the US Department of Defense, for example. One of the most interesting ideas that the comprehensive US National Security Commission on Artificial Intelligence report outlines is the need for a stronger government role in technology strategy.
Advocacy for a stronger role of government is promoted in other national AI strategies, including France’s. Here, the state is positioned as a vehicle for technological transformation: ‘The state both transforms and shows the way.’ It plays a critical role in ensuring that the public good is preserved, and reduces the trend towards monopolisation in AI.
The idea that the state could play a stronger role presents a challenge to the dominant Australian political discourse of letting the market do its thing. But this laissez-faire approach misunderstands the strong historical relationship between government and the private sector when it comes to technological innovation.
Australia’s approach to innovation, which currently centres around adoption of foreign technology, can be cheaper and faster than building a domestic capacity from the ground up. However, it’s probably time for a more mature conversation about what local industries and sectors need to be incubated, alongside which skills and labour conditions, particularly those relevant to the global race for AI talent.
The US and China certainly recognise that building critical technology capacity is not about just letting the market do its thing, although there are important differences between autocratic and democratic models of state involvement.
The recent signs from the US indicate that the Biden administration will set aside concerns about budget deficits, engage in massive spending and commit to huge investments in breakthrough technologies.
China is seeking to dominate cutting-edge technology and is also spending big, and its success in building authoritarian AI systems will result in profound challenges to democracies. Technology is geopolitical and if Australia continues to be a tech taker it will be left in a profoundly difficult position.
Despite the government’s long-standing rhetoric against deficits, the federal budget has shown signs that it is spending its way to recovery, taking advantage of record-low interest rates. In this, the government may be awakening to how strategic spending can, and must, build sovereign tech capabilities and skills. The challenge will be to ensure that the benefits of investments in AI systems and digital technologies are shared widely across Australian society.
It’s possible that Australian discussions to date have been too focused on the narrative of ‘Will a robot take my job?’. A more interesting and nuanced public conversation might be around the fact that there still so many jobs, despite increasing uptake of AI automation. This would acknowledge the complex drivers of the types and conditions of work available, and the kinds of AI industries we should be encouraging.
The French AI strategy tackles the drivers of technological development by considering the kind of AI future France wants to create in the best interests of all its citizens. This approach recognises that AI is a genuinely transformative technological development and that we must think through its implications for everything from national security to social equity and environmental and economic policy.
At the centre of these issues is the need to understand how the kinds of data we use to build AI systems can have big social and political impacts. It also means understanding how data can be either a national strategic asset or a huge vulnerability.
Without this understanding there’s no possibility of ensuring equity in AI systems. In the interests of national cohesion, Australia should ensure that AI systems we adopt here won’t re-create or exacerbate the inequalities and low productivity that have plagued Western economies and technological systems.
Cedric Villani, the author of the French AI strategy, challenges the view that successful innovation means replicating US approaches. Without strategic thinking and development of our own capabilities, much of our ability to capture value will be dependent upon the intent of those who control AI systems.
This could mean becoming a ‘cyber-colony’ dependent on others for technology and creating a society that is not devoted to humans flourishing within a robust democracy .
In a context where AI globally is being used for mass surveillance for the purposes of advertising, law enforcement and political coercion, Australia must lean into its democratic legacy. This means ensuring that where AI is being used to assist decision-making, those decisions are transparent and explainable. An Australia AI strategy also needs to buttress consumer and privacy rights as well as resist discrimination. All this may be difficult to achieve without a strong domestic AI industry and high levels of AI literacy across society and government.
Australia is a number of years behind other countries on producing a credible AI strategy, although an action plan is in the works. It aims to build on previous solid work by Standards Australia on regulation, CSIRO’s work on a tech roadmap and ethics, as well as the human rights work of the Australian Human Rights Commission. However, it unfortunately also illustrates the continued centrality of adoption in Australia’s tech policy, as well as only a $20 million average annual spend on AI research and industry–research collaboration.
Two years ago CSIRO chair David Thodey likened Australia to Kazakhstan in terms of its inability to harness technology to diversify its economy. Compared to the money being spent globally in AI research, the budget commitment of a total of $124 million over four years is tiny. But more disheartening is the absence of a truly strategic vision that ties together various strands of the digital economy and makes the case for how we should be positioning ourselves in regard to the broader geopolitical contest in the Indo-Pacific, a contest which will play out on the terrain of values as well as that of technology.
It would be a great shame if the Australian government missed this critical moment.