Some of Australia’s largest companies have signed up to trial the government’s new set of artificial intelligence principles, designed to reduce the risks associated with the technology.
NAB, Commonwealth Bank, Telstra, Microsoft and Flamingo AI will test the principles to ensure they deliver “practical benefits” as part of the federal government’s ongoing development of a national AI ethics framework.
The new ethics principles are completely voluntary and it is unclear how long the trial will last or how it will be measured.
The principles, which include things like fairness, privacy, and accountability, can be considered at any stage of an organisation’s AI technology life cycle but the government says they are not binding nor applicable for every use of AI.
Because the principles are voluntary, there are currently no consequences for the development of unethical AI technology by Australian businesses.
Australia’s leading scientists this year called for a national strategy on AI and strict regulation because of the technology’s disruptive potential.
Minister for Industry, Science and Technology Karen Andrews, whose department is responsible for Australia’s AI strategy, said yesterday she did not want to introduce a “big stick” for AI and that level of regulation “would be a long way off”, according to the ABC.
Her department, Industry, Innovation and Science, released a discussion paper in April and began a formal consultation process.
The government received over 130 submissions in less than two months from academics, citizens, and industry. It also conducted roundtables in major capital cities and collaborated with “AI experts” for a total 272 stakeholder organisations and individuals consulted.
The consultation process led to significant changes to the draft ethics principles which had been developed by the CSIRO and its digital arm, Data61.
For example, the first draft principle, “generates net benefits”, has been changed to “human, social and environmental wellbeing” after stakeholders challenged the practicality of achieving or defining a “net benefit”.
The new eight ethics principles for the trial are:
- Human, social and environmental wellbeing
- Human-centred values
- Privacy protection and security
- Reliability and safety
- Transparency and explainability
A full break down of the principles from the government is available here.
“Agreeing on these principles with business, academia and the community is a big step forward in setting our shared expectations of each other in Australia’s AI future,” Minister Andrews said in a statement.
“The government will continue to work with experts to explore the role of AI in Australia’s future and build tools to support AI development and adoption.”
Glenda Crisp, NAB chief data officer, said the bank wants to use the trial to contribute to the development of AI ethics.
“We hope to make a meaningful contribution to the discussion, to learn more about how we can leverage AI in an ethical way in order to help deliver new and improved experiences for our customers.
“Collaborating with government and across industry drives diversity of thinking which is vital in developing new ways of working and implementing new technologies safely.”
Minister Andrews’ office has been contacted for comment.