AI governance, artificial intelligence, Australia, definition of artificial intelligence, Regulating AI

Australians to develop standards for AI regulation


Standards Australia, the quasi-governmental organization for developing non-government standards down under, has released a paper calling for input on AI standards. A June 2019 discussion paper kicks off the process.

The meat of the paper starts with a “What is Artificial Intelligence?” section which is appreciated, given how many of these papers and calls for input skip over that point probably because it is a hard question. The paper adopts a definition:

“[a] collection of interrelated technologies used to solve problems autonomously and perform tasks to achieve defined objectives without explicit guidance from a human being”

Interesting to see the inclusion of an exclusion of human involvement in contrast to other definitions I have discussed. I’m not actually sure why that is a necessary element. Couldn’t AI exist with human guidance?

Highlighting that standards form an important part of a regulatory spectrum in that they promote transparency and accountability, the paper notes that both domestic and international standards will have to be addressed. Participation by the Australian body with the International Organisation for Standardization is already underway. Working groups exist to study differing AI technologies, to ensure trustworthiness (transparency, explainability, robustness, minimizing bias, reliance), and to identify domains of application. The OECD principles are noted. Appendix A to the paper is a handy cheat sheet for what is going on around the world as it lists regulatory proposals and steps taken by governments in Canada, China, the EU, Japan, Germany, the USA, Singapore, New Zealand and the UK.

Domestically, the standards body notes that there are applications and use cases in agriculture, human services, financial services, transport and logistics and mining, oil and gas. In each of these sectors, issues and opportunitses are noted. These include lack of open platforms, consent and privacy issue, data ownership, safety and interoperability.

Standards Australia wants input from industry on the following:

01 Where do you see the greatest examples, needs and
opportunities for the adoption of AI?
02 How could Australians use or apply AI now and in the future? (for example, at home and at work)
03 How can Australia best lead on AI and what do you consider
Australia’ s competitive advantage to be?
04 What extent, if at all, should standards play in providing a
practical solution for the implementation of AI? What do you
think the anticipated benefits and costs will be?
05 If standards are relevant, what should they focus on?
a) a national focus based on Australian views (i.e. Australian
b) an international focus where Australians provide input
through a voice and a vote (i.e. ISO/IEC standards)
c) any other approach
06 What do you think the focus of these standards should be?
a) Technical (interoperability, common terminology, security
b) Management systems (assurance, safety, competency etc.)
c) Governance (oversight, accountability etc.)
07 Does your organisation currently apply any de facto ‘standards’
particular to your industry or sector?
08 What are the consequences of no action in regards to AI
09 Do you have any further comments?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s