A change in Government may put the brakes on Australia’s support of AI and ADM? Or perhaps not?

The recent change to Australia’s Government will likely see a raft of changes. Of particular interest is whether and to what extent the new Labor Government will embrace artificial intelligence (or AI) and automated decision making (or ADM). You will recall the release of the former Coalition government’s Digital Economy Strategy (Strategy) in Budget 2021-2022. The Strategy set a vision for Australia to be a top 10 digital economy and society by 2030. With AI projected to contribute more than $20 trillion dollars to the global economy by 2030 and provide a potential 1.2 million new technology jobs in Australia by 2034, the Strategy firmly turned its focus to the opportunities and benefits of AI and ADM in Australia.

As the new Labor Government finds its feet, no doubt it will bring a critical lens to the existing Digital Economy Strategy with a good chance many of the proposed initiatives will be re-worked, updated or removed altogether. It is safe to say however that no matter which side of politics you are from there is no denying that support for AI and related technologies will feature strongly in Australia’s future.

Leaving aside current political dynamics, as part of positioning Australia as a leader and early adopter of AI and ADM, there has been a focus on how Australia will regulate a digital economy as new technologies become more advanced and widespread. As is always been the case the law and policy regulation cannot keep up with digital innovation. Modernising the legal frameworks and regulations is aimed at enhancing public trust and confidence around new technologies, increasing certainty around their use and benefits, mitigating risks and ultimately encouraging the increased adoption of technologies and investment in Australia’s digital economy.

In March 2022, the Digital Technology Taskforce issued the Issues Paper – Positioning Australia as a leader in digital economy regulation, with a focus on AI and ADM.

What is AI and ADM?

AI refers to a collection of interrelated technologies that can be used to solve problems or perform tasks autonomously, in some cases without human interaction. AI has the ability to learn, predict and take independent actions.

ADM refers to technology that is used to automate a decision making process. This often involves the use of rule based formulas or predictive algorithms.

Both AI and ADM are advancing quickly in Australia and overseas. AI works hard to ensure autonomous trucks safely transport iron ore to waiting driverless trains bound for Western Australian ports and waiting markets. ADM has gained enormous popularity in the financial and insurance sectors, helping to vet applications for new policies and accounts.

New opportunities

The Issues Paper emphasises the vast opportunities presented by AI and ADM to our work and personal lives. These new technologies are considered to boost productivity, improve service delivery and help solve a number of real-world problems.

The Strategy has stated that the reduction in the costs of producing and delivery goods and services will lead to lower prices and lower taxes. Environmental benefits and the ability for humans to focus on more complex and high risk tasks due to the use of AI/ADM for routine processes are also highlighted as benefits of this new technology.

In addition to the brief examples noted above, the use of AI and ADM in Australia already spans across nearly every industry and sector, including:

1. the use of AI in the medical industry to measure a patients range of motions to detect and monitor conditions such as brain aneurysm;
2. use of AI to allow the autonomous sorting of recycled materials;
3. the use of AI in agriculture to map buds, flowers and fruit counts to improve management of crops;
4. the use of ADM to automate tax returns to reduce turnaround; and
5. the use of ADM for customer service and security checks, including the biometric passport checks automated through SmartGate at Australian airports.

The state of regulation

The existing Australian Artificial Intelligence Ethic’s Framework (2019) already sets out a number of principles to follow during the AI system lifecycle. These principles are voluntary and include consideration of factors including human, societal and environmental wellbeing, privacy protection, human-centred values and accountability.

Similarly, the Automated Decision Making Better Practices Guide (2019) sets out tools and a checklist to assist in designing and implementing new automated systems. The guidance includes areas such as quality assurance processes, transparency and accountability and assessing the suitability of automated systems.

Other regulations that currently apply to new technologies include:

1. OECD/G20 AI Principles (2019) – Which promotes the use of AI that also respects human rights and decorative values.
2. Privacy Act 1988 (Cth) – The current review of the Privacy Act by the Attorney General’s Department is set to include an examination of the privacy implications of ADM with a discussion paper currently considering feedback from stakeholders. New Attorney General Mark Dreyfus has committed to ‘sweeping reforms’ during Labor’s first term in office.
3. AI Action Plan (2021) – Australia’s strategic vision to become a global leader in AI.
4. Blueprint for Critical Technologies (2021) – Identifies critical technologies that have the capacity to enhance or pose a risk to Australia’s national interest. The Blueprint includes 63 critical technologies for the national interest including AI, algorithm and hardware accelerators. The blueprint includes four goals including promoting Australia as a trusted and secure partner for investment and innovation and ensuring there is access to critical technologies that are secure and cost efficient.

Issues and concerns with AI/ADM

The Issue Paper outlines the key issues relating to AI and ADM which need to be addressed to enable Australia to succeed in its vision of becoming a global leader in this space.

The critical issues outlined in the Issues Paper include:

1. Regulatory uncertainty and complexity – Overlapping regulations, the lack of technology neutral language and contradictory compliance requirements are seen as a major deterrence to innovation and the development of technologies. The Issues Paper acknowledges that Australian regulators have been perceived as lagging behind industry in responding to new technologies, leaving developers confused regarding how they can comply with multiple sector specific regulations that do not take into account the functionality or nature of new technologies.
2. Public trust and confidence – A lack of understanding regarding how AI and ADM works is seen as a potential barrier to the uptake of new technologies in Australia. Education by the Government and private sector as well as regulation regarding risk and liability are seen as ways to provide consumers with greater certainty regarding the use of new technologies.
3. Potential for bias and transparency – A significant concern for consumers is the potential for bias and discrimination as a result of the AI/ADM reflecting the views of their programmer. Transparency in the decision making process and outcomes is seen as an option of addressing this concern. Given the complexity of algorithms used and the intellectual property and confidential information contained in technologies, there are issues regarding the practicalities and willingness of developers to divulge technical workings of their AI/ADM products.
4. Discretion – The lack of human discretion and inability to make judgment calls for bespoke scenarios is a common criticism of AI/ADM. The Issues Paper highlights that while new technologies that clearly follow a set of rules without the need for discretion may be simpler to regulate, AI and ADM which involves discretionary decision making may be more complicated. The granting of a loan, employment offer or immigration status are examples which bring this question to light. The balance between the efficiency which may be achieved through AI/ADM and the fairness of the outcome due to a lack of human discretional decision making is a topic which requires further public consideration.
5. Privacy – Technologies using AI and ADM may include the collection, examination and collation of personal information (e.g. identification information, including biometric information, financial information and health information) from various sources. While the Privacy Act 1988 (Cth) applies to these technologies, there is concern regarding whether further regulation is required to address the complexities related to such technologies.

So where to from here?

In order to meet Australia’s vision of being a global digital economy by 2030, the Digital Economy Strategy intends to focus on policies that build the right foundations that enable growth, build capabilities in new technologies and lift ambition through collaboration and strategic investment.

It is yet to be seen to what extent the new Labor Government will pick up this baton, with already a full agenda of policy initiatives competing for attention. Early signs however are positive. As noted above the Attorney General, Mark Dreyfus seems to be keen to pick up where he left off in 2013. There is talk of a new tort for serious breaches of privacy and alike and he has also announced he will publish the proposed privacy reforms currently being considered. Whether this will translate into substantive policy reform and the necessary investment to encourage, incentivise and fund digital innovation, including AI and ADM is just not clear at this point. Let’s hope there is clarity soon or Australia will be left behind at a critical juncture. Neither AI nor ADM innovation has the patience to wait.

If you found this insight article useful and you would like to subscribe to Gadens’ updates, click here.

Authored by: 

Dudley Kneller, Partner
Eve Lillas, Associate