Artificial intelligence, in its many forms, is rapidly embedding itself in and transforming many aspects of our everyday lives, often without our knowledge. This means that AI has the potential to fundamentally change how we live for the better — but only if its implications are well understood and its potential risks are mitigated.
Governments around the world are signalling their commitment to AI and their role in shaping the AI agenda of the future through the release of national AI strategies and frameworks.1 On Friday, 18 June 2021, the Australian Government joined them in releasing Australia’s AI Action Plan (the Action Plan), outlining its vision to ‘establish Australia as a global leader in developing and adopting trusted, secure and responsible AI’. The Action Plan sets out four key areas of focus, and associated policy measures, to help realise this vision:
- Developing and adopting AI to transform Australian businesses.
- Creating an environment to grow and attract the world’s best AI talent.
- Using cutting edge AI technologies to solve Australia’s national challenges.
- Making Australia a global leader in responsible and inclusive AI.
The release of the Action Plan follows hot on the heels of the Australian Human Rights Commission’s Human Rights and Technology Final Report (the AHRC Report), which set out a comprehensive analysis of the current landscape and proposed a roadmap for Australia to seize the opportunities of AI and address the potential threats that it poses to human rights. The Action Plan also reiterates the government’s commitment to consider the AHRC Report.
What these recent developments make clear is that the promise (and potential peril) of AI is no longer theoretical, and it is well and truly time to take active steps to harness it. However, it is not clear whether the policy measures set forth in the Action Plan — particularly in relation to the fourth focus area, ‘Making Australia a global leader in responsible and inclusive AI’ — will be enough, without more, to do so.
What are the challenges involved with the rising use of AI?
In our response to the Government’s earlier call for views on the Action Plan, we strongly supported the creation of the Action Plan, and outlined why it was so urgently necessary.
We believe that for AI to create real social value, it must be designed, adopted and used in ways that ensure its trustworthiness in order to build and sustain public confidence. The rapid pace at which the economy is digitising makes this challenging from a number of perspectives, including those identified below.
How do we holistically assess harms and benefits of AI?
It is not clear that any one suite of policy measures will be sufficient to address these challenges. In our response, we proposed a flexible methodology in order for the Australian Government (and others) to assess and determine how to pragmatically respond to the impacts and opportunities associated with AI on an ongoing basis.
We saw this approach consisting of three key stages:
- Understanding the use cases.
First, we need to understand how AI is being used. Who is, or could be, using or impacted by the AI solution? In what context is it being used? How may it be used in different stages of its lifecycle? How does it interoperate and interconnect with other AI solutions or emerging technologies?
- Assessing the use cases against a framework of overarching principles
Once we have established the use case for an AI solution, we suggested that it be assessed against a framework to determine the nature and extent of any actual or potential harms or benefits in the relevant use case(s) — and the likelihood that they will eventuate.
- Prioritising and evaluating potential responses
Once potential harms and benefits have been identified, we suggested a final stage of considering how to prioritise and determine what actions to take (if any) to ensure that the benefits are maximised, or the harms are minimised. This should take into account existing laws and inform policy initiatives designed to minimise the impact of the ‘digital divide’ across society.
This approach is summarised below.
Where to from here?
The Action Plan outlines a number of steps that will be taken in the immediate term, including consideration of the Australian Human Rights Commission’s detailed recommendations in the AHRC Report, and continued commitment to the implementation of Australia’s AI Ethics Principles.
Ultimately, the measures set forth in the Action Plan are, necessarily, just the beginning. The Action Plan and the AHRC Report do not only reflect increasing societal awareness of the need for responsible, ethical AI — they also illustrate the increasing need for concrete action to achieve it.
For further information, please download Herbert Smith Freehills’ response to the Australian Government’s Discussion Paper on the AI Action Plan.