Artificial intelligence is already ubiquitous in our lives and will be foundational to our future prosperity. In Australia, the Department of the Prime Minister and Cabinet has identified AI and machine learning as critical technologies of national interest. AI applications are wide-ranging and encompass many sectors, including agriculture, education, energy, finance, health care, manufacturing, transport and telecommunications.

As a general-purpose technology, AI applications in those domains may also have implications for defence and security. Indeed, the Australian Defence Force’s ability to understand the operational environment, manoeuvre and project force will be transformed through advances in AI and human–machine partnerships, employing dual-use and defence-specific AI applications.

Innovation in AI is accelerating, creating opportunities to grow the economy, solve challenging problems, accelerate discovery and enhance national security. It’s estimated that AI could exceed $20 trillion in worth to the global economy by 2030. This potential for transformative impact, and the competitive advantages offered by AI, are being actively pursued around the world. More than 30 countries and regions had published AI strategies or similar documents by December 2020; the world’s top universities are increasing their investment in AI education; and private investment in AI continues to grow. This ongoing expansion in AI activity will bring accelerating advances across the field.

With global AI investments growing, Australia will need to remain internationally competitive in a world of expanding AI capabilities and AI-enabled operations. For defence, AI will be critical in delivering strategic objectives and maintaining a capable, agile and potent defence force. Reflecting that importance, and the need for increased self-reliance, AI is included among the new ‘sovereign industry capability priorities’ critical to defence. A sovereign capability in AI ensures Australia’s stake in a key 21st-century industry and underpins the defence organisation’s future operational and training capabilities.

Australia has some world-leading AI capabilities in universities, research organisations and industry. To expand from that base, we need to focus on AI and broader STEM (science, technology, engineering and mathematics) initiatives to build the specialist workforce required by Australian industry. Building human capacity in AI will be critical, as demand is already outstripping supply, and the need for AI specialist skills is expected to grow significantly. For sovereign activities, including sensitive work supporting the nation’s defence and security sector, the ability to draw upon an Australian AI-skilled workforce will be fundamental.

Capacity in the Australian workforce will ensure that we can maintain an effective defence force and develop the capabilities needed to meet and overcome the disruption that AI will introduce to current methods of warfare. That will require sustained investment in skills training at all levels to raise understanding of AI by managers and to further expand AI graduate and higher degree by research programs to attract and train Australian AI specialists.

Increased research and technology development, supported by long-term and sustained investments in data, tools and critical digital infrastructure, will underpin our national AI competitiveness. Through sovereign investment in AI research and development, the benefits of AI will be fully harnessed and the potential risks mitigated. Investment should include building access to sufficient computing capacity and the fast networking needed to source those capabilities.

In addition, data holdings need to be established and managed to support the development of AI applications and tools, buttressed by appropriate access and privacy protections. Issues of safety, security and bias need to be understood and actively researched to support and inform responsible AI practices and build trust in AI systems to underpin confidence in continued AI development. That will require the development of AI systems that accord with Australian values and are trusted partners across the breadth of AI applications. The ongoing development of AI standards, engaging a broad cross-section of stakeholders, will enable the widespread use of responsible AI and increase the competitiveness of Australian industry.

Through government leadership, a strong and vibrant AI ecosystem should be established to bring together partnerships with industry and academia to drive innovation, meet national priorities and expand AI exports. This will ensure strong stakeholder participation and dialogue to inform policy considerations and connect industry and academia with new business opportunities. Efforts are already being made in that direction. In the defence sector, this includes the establishment of the Defence Artificial Intelligence Research Network. The DAIRNet will work to build a research community in Australia supporting Defence, bringing together multidisciplinary teams to address AI challenges at scale. Beyond research and innovation, a sovereign capability entails the adoption and continuous upgrade of AI technologies to enhance productivity, requiring industry participation across the AI product life cycle.

This article is an extract from a new report, Artificial intelligence: your questions answered, produced by ASPI in conjunction with the Australian Institute for Machine Learning at the University of Adelaide.

Last week, Prime Minister Scott Morrison announced an increase of 18,500 to the Department of Defence’s full-time workforce by 2040. I’ll spare you the (semi-)snarky commentary about (semi-)announcements this time and get straight to unpacking what we know about this measure.

The government hasn’t released a detailed breakdown of what the 18,500 will be doing, but it did provide some high-level information to the media. From that, it appears there will be two phases of growth. Over the next decade there will be 12,500 new positions. Those have already been assigned to particular new capabilities. In the second decade there are 6,000 more positions, which seem to be more of a general pool that will be assigned to roles as Defence’s long-term requirements are better understood.

Around 2,000 of the first 12,500 are Australian public service positions, with another 2,000 in the following 6,000. That’s understood to be in addition to the roughly 540 new APS positions announced in Defence’s mid-year budget update that are intended to help deliver AUKUS programs and the sovereign guided weapons enterprise. All up, that’s a 15% increase to Defence’s APS numbers. Since more public servants generally isn’t a vote-winner outside the nation’s capital, the government hasn’t been playing this up. But it suggests the government has realised Defence can’t deliver its huge investment program without people to run acquisition projects. Some of those new positions are also going to be in the Australian Signals Directorate for cyber capability.

That leaves around 14,500 new full-time uniformed positions in the Australian Defence Force. Considering the ADF’s strength is currently a little over 60,000, that’s a very substantial increase of around 25%. That prompts two key questions: what are they for and can Defence actually recruit them?

There’s a simple answer to the first question. They are needed to operate the new capabilities that are already in Defence’s investment plan, the 2020 force structure plan that accompanied the 2020 defence strategic update (which ASPI has analysed in some detail). It gave Defence around 1,000 additional people as an immediate measure but stated, ‘A detailed proposal for … longer term growth will be considered by Government in 2021.’

Since then, Defence’s workforce planners have been hard at work mapping people to those planned capabilities. The aggregated answer is the 18,500 additional people, which the government has now considered and agreed to. The $38 billion to pay for those people over the next decade is already built into Defence’s 10-year funding model set out in the 2020 update.

A cynic might say there’s little new in the recent announcement. But it’s important to align capability with people and the government has now done that. There’s no point acquiring military systems if you don’t have the people to operate them effectively.

Those people are needed to increase the number of submariners from 800 to potentially more than 2,000, to operate new land-based strike capabilities and to conduct cyber and information warfare, among many other tasks associated with new capabilities.

What Australians won’t see is a new, dedicated emergency response function. Despite the uninterrupted chain of civil emergencies over the past several years, it looks like Defence is still planning to meet those contingencies with its core warfighting capabilities. Whether that makes sense is a separate conversation.

The second question is whether Defence will be able to recruit those people. We’re all familiar with stories of Defence’s struggles to recruit and retain people. And recent history isn’t particularly encouraging. The ADF has only grown by about 2,000 people since the 2016 defence white paper, or fewer than 400 per year on average. It now has to sustain average annual growth of twice that for 20 years.

But there are also grounds for optimism. The current force is only 0.23% of Australia’s population—less than one-quarter of one percent, or only one in every 400 Australians. While a future full-time ADF strength of 75,000 to 80,000 sounds like a lot, it will still be only a very small percentage of Australia’s population. Considering that our population has grown by around 34% over the past 20 years and neither major political party seems to want to fundamentally change our immigration policy, it’s reasonable to assume future population growth will be similar. Put another way, the ADF just has to preserve its current share of the Australian population to achieve the target.

However, the issue may not be numbers of people per se but developing the skills required for the new capabilities. Some of the new positions will be for people with skills the ADF currently doesn’t have and who will take years to train (like nuclear-qualified submariners). Others require skills that are highly sought after and can attract much higher remuneration in the private sector (like cyber experts). The hit to migration wrought by the Covid-19 pandemic certainly isn’t helping as it has increased competition for skilled people.

The key may not be in recruitment, but in retention. The ADF has averaged a separation rate of 8–10% over recent years. In essence, that means it’s losing 5,000–6,000 people per year, many of whom are already trained and highly skilled. If it can reduce that by 1,000 per year, it will have basically cracked the problem. Of course, that’s easier said than done. It comes down to giving ADF members a competitive employment offer. Defence is aware of this and working on it. The business case for an east coast submarine base is as much about recruitment and retention as it is about the strategic drivers of a two-ocean navy. And the ADF’s total workforce system seeks to provide its members with some of the flexibility that Australian workers now take for granted.

The question isn’t so much about whether the growth is achievable; it’s about whether a defence organisation that is used to developing and working to 20-year plans can meet the threats posed by an uncertain world that’s rapidly changing. The 2020 update rightly said we can no longer count on 10 years of warning time. The day Russia’s Vladimir Putin invaded Ukraine, many pundits and politicians were still saying it wasn’t going to happen. And now we face the prospect of China’s Xi Jinping exploiting war in Europe to further his own designs of remaking the regional and global order.

Aligning workforce requirements with capabilities is a good thing. But if we aren’t getting the people we need for 20 years, that means we won’t be getting the capability we need for 20 years either.

Australia’s role in critical space domain awareness (SDA) is becoming a major mission for the defence force. Australia has long provided a ‘suitable piece of real estate’ as part of the Five Eyes intelligence-sharing community and through the 2014 Combined Space Operations initiative. The continent lies directly beneath launch trajectories from Chinese launch sites and is well located to monitor activities out to geosynchronous equatorial orbit (GEO), 35,786 kilometres from earth. Inland, the sparsely populated landmass offers skies largely free from light pollution, and while weather will always be a factor—ground-based optical telescopes can’t see through cloud—Australia is better placed to undertake the SDA mission than most other locations.

The United States has relocated a C-band space surveillance radar and a large space surveillance telescope to the Northwest Cape region near Exmouth in Western Australia as a hosted joint facility that will be operational this year. SDA is part of the Australian Defence Force’s Operation Dyurra, which integrates space capabilities, services and effects into wider operations, and is further integrated into the US-led Operation Olympic Defender providing global space surveillance of potential threats within America’s space surveillance network.

Australia’s Defence Department is seeking to expand sovereign SDA capabilities with a major project, JP-9360, and released an SDA roadmap in July 2020. That plan consolidates individual projects to provide a more cohesive and connected but distributed and multilayered approach to SDA. The first tranches will be delivery and introduction of a telescope.

In tranche 2, a ground-based network of optical sensors will be deployed across Australia, with different types of sensors providing enhanced and expanded space surveillance beyond the Exmouth-based telescope. Approval for this work is expected in late 2022. An example of what it might include is the laser-optical system demonstrated by Australian company EOS Australia from facilities at Mt Stromlo near Canberra and from Learmonth in Western Australia. Lockheed Martin, in partnership with Curtin University, has demonstrated the FireOPAL sensor network, offering a low-cost distributed system for comprehensive ground-based space surveillance.

What’s next? A logical approach would be to complement ground-based optical sensors with additional systems operating in other wavelengths, such as radar or infrared, and then locally develop and deploy space-based space surveillance satellites.

Space-based space surveillance offers advantages over ground-based sensors. Apart from some specialised technologies, most ground-based sensors, particularly those operating in the optical part of the electromagnetic spectrum, don’t work very well in daylight so surveillance is limited to dusk, nighttime and just before dawn. And, again, weather can impede the effectiveness of ground-based optical sensors, even at nighttime.

Satellites orbit beyond the effects of terrestrial weather, and in space there’s no daytime or nighttime to interfere with surveillance. From orbit, a satellite can observe a target of interest or survey activity in its field of view across a range of orbits and trajectories. Adelaide-based Inovor Technologies is already developing small satellites under Project Hyperion that will be based in low-earth orbit to observe satellites, space debris and other activities in medium-earth orbit (MEO) and out to GEO.

Inovor notes that its satellites can observe any GEO object from multiple vantage points simultaneously to help identify it. Hyperion’s ability to have ‘persistent custody’ of any GEO or MEO object of interest will improve observation of the space environment and monitoring of suspicious activity and will better support missions such as a rendezvous for servicing in orbit.

The potential offered by space-based space surveillance also extends to close approach operations to examine a target in detail. Such rendezvous and proximity operations need to be managed carefully and carried out consistent with space law. International diplomatic efforts are underway to define what constitutes responsible behaviour in space, including in rendezvous and proximity operations. Australia is fully engaged in this dialogue via a UN open-ended working group established late last year following the tabling of UN General Assembly resolution 75-36 by the United Kingdom.

Australia’s approach to SDA needs to be consistent with international law, so care must be taken in shaping ADF capability to ensure all activities meet standards of responsible behaviour. The increasingly complex and congested nature of space and the real challenge posed by adversary counterspace capabilities mean that Australia has an international responsibility to contribute to effective SDA. Moving from a purely ground-based system to a more sophisticated combination of space-based and ground-based space surveillance capabilities makes eminent sense. By taking this approach, the ADF will be best placed to contribute towards effective SDA as part of the Combined Space Operations initiative and, more broadly, at an international level to ensure responsible behaviour in space.

The new Defence Space Command should take an ambitious and forward-looking approach to the SDA mission, and it needs to move relatively quickly to field new capabilities. Space is an area in which force development and capability acquisition can occur relatively quickly, especially if the government is willing to support small and medium enterprises in Australia’s commercial space sector.

SDA opens real potential for Australian companies to contribute valuable sovereign capability for the ADF that doesn’t necessarily rely on overseas primes and give added momentum to the growth of Australia’s space sector. It’s important for the government to keep looking up, and to help Australian businesses provide capability solutions for vital defence tasks.

The Department of Defence has exceeded its targets for contracts awarded to Indigenous Australian businesses by 600% in the latest data release on the government’s Indigenous procurement policy. Although this represents a doubling of contract value, from $300 million to $600 million, a House of Representatives committee report tabled in August suggests that consolidating the policy would push procurement towards further quality alongside volume.

The Indigenous procurement policy has been a key way to stimulate the development of First Nations enterprises by setting annual targets for the number and value of contracts awarded to Indigenous businesses in each federal government portfolio.

When the policy was introduced in 2015, Defence aimed to award 70 contracts to Indigenous enterprises; instead, it delivered 278 contracts amounting to $159.3 million that year.

In 2019–20, the department outstripped its target of 676 contracts, awarding 6,476 contracts to Indigenous businesses worth $610 million. Defence was by far the largest procurer of Indigenous business, followed by Health ($80 million) and Social Services ($72 million).

The Indigenous procurement policy has been a bright spot in the wider policy area of Indigenous affairs, where the complexity of vested interests, stigma and discrimination ensure that inertia is common and even minor changes are tremendously contentious.

There has been a flurry of announcements of Indigenous policy over the past few months. But, unfortunately, a common characteristic of these initiatives is the deferral of actual delivery on many programs. It’s a worrying sign that Indigenous economic and social development continues to be a second-tier priority.

The government announced in August that it would work towards a national roadmap for Indigenous skills, jobs and wealth creation, which would aim to provide a long-term commitment and organising framework to implement actions that increase economic opportunities for Indigenous Australians. It’s unclear, however, whether and how this is different from the Indigenous Reference Group on Northern Australia (whose recommendations to government have not been made public) or the Closing the Gap initiative, which was already meant to provide a framework for change.

In September, the government released a discussion paper on developing a digital inclusion plan, which would act on the triennial regional telecommunications inquiry that, since 2008, has identified an acute need for better telecommunications services for those living in rural and remote Australia.

In this context, Defence’s achievement, which involves actual delivery at levels beyond its commitments, is nothing to be sneezed at. With exponential growth in contract value, there are signs that Defence’s momentum won’t lose pace. Nonetheless, it’s important to consider ways in which its current commitment can be strengthened.

Earlier this year, the House of Representatives Standing Committee on Indigenous Affairs heard testimony from numerous First Nations people to understand the opportunities for economic development and employment for Indigenous Australians as well as the barriers.

The committee heard testimony on a host of issues, including about ‘black cladding’ (creating a management structure that satisfies the ownership criteria of the Indigenous procurement policy but where control of the enterprise may be vested in the hands of non-Indigenous managers). The committee recommended a series of random, independent audits of entities that have been awarded contracts to ensure that black cladding isn’t happening.

The committee recommended an investigation of how Indigenous businesses are defined under the policy. This would include consideration of alternative measures such as the proportion of Aboriginal and Torres Strait Islander people employed, how skills and training are transferred and how company profits are used to promote philanthropic causes. Many Indigenous businesses have sponsored young people through education pathways, including Baidam, which has funded cybersecurity qualifications, and Australian Expedition Vehicles, which has placed students from the Proud Warrior program.

Although the overall volume of contracts awarded to Indigenous enterprises is high, the average value of contracts is low compared with non-Indigenous businesses. This is a sign that Indigenous businesses need opportunities in business development and capacity building. The committee recommended expansion of Indigenous business hubs and business networking and of financing of Aboriginal and Torres Strait Islander housing and business loans. There are also opportunities in international trade and investment for the expansion of the Aboriginal and Torres Strait Islander business sector.

Wider issues include a ‘data shortfall’ in tracking the role of Indigenous businesses in providing employment. The Central Analytics Hub in the Department of the Prime Minister and Cabinet initiated a project in 2019 to improve the quality of labour force statistics for Indigenous Australians. However, the project was shelved because of the pandemic and bushfire season. When contacted by ASPI, the department reported that the feasibility of the project was limited by significant issues with data access.

Notwithstanding the many challenges of making Indigenous affairs part of the core business of Defence, there’s widespread senior executive support and an overall commitment to use scale effects to lift Indigenous communities via the business sector. This is largely a reflection of the fact that Indigenous businesses have substantially better employment outcomes than non-Indigenous businesses. The House Indigenous Affairs Committee’s inquiry found that the employment rate for Aboriginal and Torres Strait Islander people in Indigenous businesses is 60% higher than in other businesses.

Testimony to the committee showcased the ways in which Indigenous business owners are absolute experts in business and are building their capabilities while at the same time making a difference in communities by providing career options and second chances to young people. With some concerted government support, the next phase of Indigenous procurement is bright.

ASPI’s International Cyber Policy Centre recently hosted a roundtable on the future of the Indigenous procurement policy, featuring the recently appointed Defence Indigenous Champion Celia Perkins and a panel discussion with Australian Army Elder Lorraine Hatton, army veteran Edwin Jim Mi Mi, Bullroarers Co-Founder and Director Neal McGarrity, Michael McMillan of Australian Expedition Vehicles, and Darren Godwell, CEO of i2i Global.

In my last post, I explained the need for dual lines of responsibility for nuclear safety, with civilian authorities answerable to a minister other than the minister for defence to provide a degree of independence in discharging the government’s nuclear safety obligations. The Royal Australian Navy and Department of Defence are answerable to the defence minister.

The submarine’s crew is the obvious starting point. A British Astute-class nuclear-powered attack submarine (SSN) crew of 98 is almost double that of Australia’s Collins-class conventional submarines. We will need to train about 30 relief crew members to ensure we can sustain each submarine with a complement of at least 130 nuclear-qualified personnel. New skills are required for the nuclear technicians operating the reactor and for steam plant operators.

The mix of skills is very different. The SSN has two command qualified officers and five marine engineering officer and the Collins has one of each. Both are scarce categories in the Royal Australian Navy and in Britain’s Royal Navy. All will require nuclear training and sea experience and will need to join some time before their submarine is commissioned. A group of 35 engineers will have to join three years before commissioning.

The RAN submarine squadron based at HMAS Stirling provides the first line of technical support, maintenance assistance and oversight for the submarine. Accordingly, the squadron is a key component in the nuclear safety chain, providing oversight of the onboard qualification process at every step in individual and team training.

The squadron conducts many of the assessment boards, has oversight of training standards and conducts the operational readiness assessments of each crew prior to certifying the submarine as ready for operational deployment. Using the RN’s practice as a guide, the number of personnel in the RAN squadron will have to increase from 109 to over 200. Of that 100 additional personnel, about 50 will be required at the start of phase 1 of the transition, two years before the submarine is commissioned and when the reactor is first activated with its ‘initial criticality and power range testing’.

Monitoring and protecting the health of the nuclear power plant operators and ensuring that the public faces no risk of exposure to harmful ionising radiation are responsibilities of those owning a nuclear power plant. The RAN radiological protection and incident response team will undertake the routine radiation monitoring function, maintaining records of the radiation dose received by watch keepers and workers to ensure that no individual is put at risk.

The team must have the expertise to respond to any nuclear incident, monitoring radiation levels and hazards to the public. While it’s to be hoped this expertise will never be applied in earnest, it requires skills that must be both maintained and exercised and the capability to be publicly demonstrated. Three teams of 12 personnel (a total of 36) will be required to ensure that a team is available around the clock to respond to an incident at the SSN home port in HMAS Stirling and at the shipyard in Adelaide. The Stirling team would be set up to travel in support of port visits elsewhere. These teams must be in place, trained and certified prior to commissioning the reactor in the first SSN, 18 months before the submarine is commissioned.

The naval nuclear training and safety panel should be a Canberra-based RAN policy directorate, headed by a captain with experience as the submarine squadron chief engineer. The organisation’s role is to support the defence minister in discharging his or her responsibility to ensure that the standards of training and qualification of nuclear power plant operators are maintained. This entails oversight of the training delivered by the RAN and any other training provider. The organisation has access to all aspects of the training pipeline for scrutiny and assessment. A team of 20 is envisaged and 16 are required in phase 1.

The submarine capability branch, led by a rear admiral, should be responsible to the chief of navy for the safe and effective delivery of the agreed level of submarine capability, within allocated resources. Operating in a matrix management structure reaching across the Defence Department and the RAN, this officer should have authority over all Defence elements involved.

This organisation would oversee reactor safety, similar to the US Navy’s operational reactor safeguards examination program, including orchestrating the unannounced drills and inspections required across the entire SSN program to ensure that this critical safety system is functioning correctly. An additional 110 personnel will be required and 50 will need to be in place during the policy development phase of the project, prior to phase 1.

The supplier navy’s submarine design authority will retain responsibility for nuclear safety within the design. A RAN liaison team will provide an interface with the design authority and a conduit to the RAN and its major maintenance facilities, with qualified personnel based in the supplier navy’s design authority and in Australia. The team will also be able to analyse any incidents and provide instructions to prevent repetition. That could involve changes in personnel skills, training, operating procedures, materials or design. A team of 25 uniformed and civilian professionals is envisaged.

Although the submarine should not require refuelling, it will require maintenance work on the reactor and its supporting systems. This work will be undertaken by two nuclear repair facility teams. If we were to follow UK procedures, team 1 would consist of 107 uniformed personnel sited at HMAS Stirling. Team 2, with 70 civilian and 57 uniformed personnel, would be situated at the shipyard in Adelaide. In the event of an incident, these personnel would assist the radiological response teams. Team 1 would need to have 25 of its personnel prior to phase 1.

In my next post, I’ll consider the training infrastructure and sum up the total number of personnel required to crew and safely oversee our first SSN.