Unexplained delays to the launch of the National Disability Insurance Agency’s (NDIA) AI-driven virtual assistant, Nadia, have raised questions about whether some government agencies have become gun-shy over recent embarrassing tech-issues relating to the ATO, Centrelink and the Census.
The NDIA’s online chatbot, for which actress Cate Blanchett loaned her voice, has been developed as the face of the National Disability Insurance Scheme (NDIS) and was due to begin a 12-month learning period in the middle of this year.
This is a critical step towards enabling Nadia to acquire the knowledge and insights she will need to answer the questions of a predicted 460,000 people being transitioned into the scheme over the next three years.
Back in February, NDIA deputy chief, Louise Glanville said Nadia could already understand and respond to thousands of questions, but needed to engage with users to expand her capabilities.
“It will take 12 months and a great deal of interactions with NDIS stakeholders for Nadia to become fully operational,” Glanville said at the time. “The more interactions she has with people, the more her knowledge bank will grow.”
Little has been said about the NDIS’ reasons for delaying Nadia’s rollout. However, it’s important to recognise that a learning AI is like a child that requires tutelage by being exposed to carefully curated data over a period of time to shape its knowledge and responses before it can be expected to provide a robust and reliable service.
I congratulate the NDIS’ pioneering work in developing Nadia to enhance its service delivery and encourage ongoing tutelage to prepare the chatbot for its eventual public release.
If Australia is to play a leadership role in developing and applying new technologies like AI, we must be willing to innovate within carefully defined safe settings and boundaries. The alternative is to see our best minds and innovation head overseas where the appetite might be greater.
One way to minimise risks when working with AI systems is to apply safe data frameworks to provide a “protected environment” within which systems like Nadia can be nurtured until they are proven ready to interact with the wider community.
Until now, little work has been done to define standards and usage principles for data frameworks. However, the release last week by ACS of a Data Sharing Frameworks technical white paper offers a new level of clarity and direction for organisations wanting to develop and use data for smart services.
The NSW Government, already the leader in open data initiatives in Australia, has been working with ACS and other industry stakeholders for more than a year as part of a Data Taskforce, seeking to identify and clarify best practice data sharing frameworks.
The taskforce, which is jointly led by ACS and the NSW Government Data Analytics Centre (DAC), is focused on several key areas including:
· cross-jurisdictional open data sharing between governments;
· data governance;
· privacy protection;
· maintaining trust of consumers and citizens; and
· safe and practical data sharing.
To ensure the relevance and value of this initiative, ACS gathered together a diverse group of stakeholders including representatives from several government agencies and jurisdictions along with key industry groups and business leaders.
We were privileged to have both the Federal Treasurer the Hon Scott Morrison and NSW Finance Minister Victor Dominello on hand to participate in the launch of the Data Sharing Frameworks paper last Thursday morning in Sydney.
The paper, which explores a range of models to support data sharing across government and the private sector, includes several recommendations designed to provide clarity and direction for best practice. You can find the full paper and its recommendations here.
It’s important to recognise that raw data has only limited value in its own right. Rather, the value of data can be increased – and even multiplied – depending on the context in which it is collected, used, and aggregated, and the range of services to which it contributes.
The ability of governments to open up datasets that have been appropriately anonymised, and make them available for use in a range of commercial and public sector applications, has the potential to drive tremendous progress in areas such as health, financial services, retail, government service delivery, future planning, education, transport and more.
The fact that data can be reused in different ways and even tracked over time further enhances its value. So does its ability to create access to new and emerging markets, provide real-time insights into financial market movements and inform predictions about future trends.
The work of the ACS Data Taskforce provides a framework for agencies and businesses seeking to leverage the growing flood of data to facilitate the creation of smart services.
I wish Nadia and NDIS all the success in navigating and embracing the brave new world.