In brief
On 18 March 2022, the government announced that it has opened consultation on digital economy regulation settings in Australia. The government is seeking feedback on regulatory frameworks, particularly in relation to automated decision making (‘ADM’) and artificial intelligence (‘AI’).
Contents
In more detail
The main issues the government is seeking feedback on include:
- uncertainty and complexity – there is regulatory uncertainty for both industry and government, as they are subject to multiple regulatory frameworks. Complexity in the regulatory environment deters innovation within government and industry. However, siloed and regime specific regulatory approaches have intensified problems of overlapping regulating and created barriers to the development and effective use of AI and ADM;
- rapidly evolving international developments – Australia faces strong global competition when seeking to be a lead in digital economy regulation. Regulation of AI and ADM in Australia needs to keep pace with these developments, whilst taking into account development sin Australia’s key trading partners to ensure consistency;
- public trust and confidence – public trust and confidence in AI and ADM is a potential barrier to achieving the enormous benefits that they can offer. There needs to be a responsibility for improving public understanding of these technologies, and greater guidance provided to users of these technologies;
- potential for bias or discrimination – it is recognized that while automation has the potential to reduce bias and discrimination, algorithms may reflect the bias of their programmers. The potential for bias or discrimination is a significant issue for the designers of systems implementing AI or ADM;
- transparency and ability to explain decision or outcomes – there may be issues posed with ensuring transparency with systems that use ADM or AI. This is mainly due to the algorithms and technological language used by machines being largely inaccessible and unable to be understood by the public. Conversely, it is unclear the extent of transparency that should be required, as users of AI or ADM are unwilling to share the details of how their relevant systems work;
- exercise of discretion – discretion is often required to ensure fair outcomes when rigid application of a rule could lead to an unintended outcome. However, regulation by simple rules without discretion is easier to automate. Therefore, a balance needs to be struck between efficiency and fairness; and
- privacy – personal information is regularly used in AI and ADM systems across a range of sectors. The role and scope of privacy laws in relation to ADM and AI is therefore important, and may need to be expanded.
The consultation poses 10 questions that relate to the above issues, and seeks feedback that will assist overcoming barriers to achieving ADM and AI potential. The consultation closes on 22 April 2022. Further information about the issue paper can be found here.
This article was originally published in the March 2022 edition of LegalBytes, which can be found here.
