AI roadmap aims to put Australia at the forefront of shaping international standards

Provides key recommendations for government, industry and other stakeholders

ai artificial intelligence ml machine learning robot touch human hand
kentoh / Getty Images

Standards Australia has published the Artificial Intelligence Standards Roadmap: Making Australia’s Voice Heard, which was commissioned by the Department of Industry, Science, Energy and Resources and aims to place Australia at the forefront of the creation of international AI standards.

The roadmap was created following a wide-ranging consultation, seeking input from a range of stakeholders in industry, government, civil society and academia.

It has eight recommendations on where Australia should focus on AI Standards work and who should have responsibility for their implementation, with recommendations being divided by four goals. These range from safety and trust for citizens and consumers, to ensuring export opportunities for Australian businesses are protected, promoted and enhanced.

The goals of Australia’s AI standards

The first goal is to ensure that Australia can effectively influence AI standards development globally. To achieve that, the roadmap recommends increasing the membership of the Artificial Intelligence Standards Mirror Committee — established by Standards Australia in late 2018 (IT-043) — to include participation from more sectors of the economy and society. The roadmap timeline is to be completed by March 2020.

It also recommends exploring enhanced cooperation with the United States National Institute for Standards and Technology (NIST) and other standards development organisations (SDOs) in order to enhance Australia’s knowledge and influence in international AI Standards development, also suggested for completion this month. 

The Australian Government nominates experts to participate in ISO/IEC/JTC 1/SC 42 and the National Mirror Committee (IT-043). The Australian Government should also fund and support their participation, particularly at international decision-making meetings where key decisions are made, within existing budgetary means. This action is expected to be completed by April.

The next goal is to increase international competitiveness for Australian businesses for AI and also streamline requirements in areas like privacy risk management.

To achieve this, Australian businesses and government agencies must work together to develop a proposal for a direct text adoption of ISO/IEC 27701 (Privacy Information Management), with an annex mapped to local Australian privacy law requirements. 

“This will provide Australian businesses and the community with improved privacy risk management frameworks that align with local requirements and potentially those of the GDPR, CBPR and other regional privacy frameworks,” stated the roadmap, recommending it be completed by June.

Another goal is to ensure AI-related standards are developed in a way that takes into account diversity and inclusion, ensures fairness and builds social trust. 

Measures should see Australian Government stakeholders, combined with industry input, develop a proposal to improve data quality in public services to optimise decision-making, minimise bias and error and improve citizen interactions. This is expected to be completed by November.

It also recommends that Australian stakeholders channel their concerns about inclusion, through participating in the Standards Australia AI Committee, and actively shape the development of an international management system standard for AI as a pathway to certification, suggested to be done by next month.

And the last goal is to grow Australia’s capacity to develop and share best practice in the design, deployment and evaluation of AI systems.

The roadmap recommends that the Australian Government considers supporting the development of a security-by-design initiative, which leverages existing standards used in the market, and which recognises and supports the work being carried out by Australia’s safety by-design initiative, suggested to be completed by December.

But before this, it recommends the development of a proposal for a standards hub setup to improve collaboration between standards-setters, industry certification bodies and industry participants, to trial new more agile approaches to AI Standards for Australia. This is expected for June 2020.

“This roadmap aims to give the Australian market a better chance of capitalising on the growing global AI sector,” Standards Australia CEO Adrian O’Connell said. “The clear and actionable recommendations in the report will support Australia to reach its full potential in the adoption and use of this technology.”

“AI is an exciting technology with a growing future in the Australian and global market. Through standards we believe we can help build confidence and safe-guard against the irresponsible use of this technology and its data,” continued O’Connell.

How standards can shape a better AI

Standards can play a key role not only in the responsible use of AI but also by establishing common building blocks and risk management frameworks that can benefit users and facilitate the commercialisation internally and internationally of AI products.

The roadmap defines three ways standards can have an impact: on a voluntary basis, by  outlining controls for information security and privacy, an auditing framework or reporting templates; through inclusion in contracts, by establishing common parameters between partners in AI deployment; and via a regulatory call-up in specific sectors in relation to medical devices or financial services, for example.

“Standards can also help protect Australia’s national and international interests. For example, they can help in implementing the Organisation for Economic Co-operation and Development (OECD) Principles on Artificial Intelligence, which Australia has supported,” according to the roadmap. “Here, standards can lead to the creation of risk management approaches, as well as more granular technical solutions that provide guidance on these principles within companies, within customer bases and across the broader community.”

There is also an opportunity for Australia to move from a nation taking standards to one making standards, which the roadmap suggests can be achieved by strengthening participation through international Standards Development Organisation (SDOs).

“There are practical reasons for this. Australia is an outwardly-facing trade-dependent nation. To responsibly share, and protect, our ideas, our products and our services, we need to act in a strategic and considered way globally. This should promote free and fair trade and enhance our national interest, including security considerations.”

Copyright © 2020 IDG Communications, Inc.

  
Shop Tech Products at Amazon