A new government taskforce has been tasked with forging a set of guardrails for the use of artificial intelligence in the Australian Public Service, as the Albanese government seeks to prevent a repeat of a robodebt-style automation fiasco.
The creation of the taskforce comes as the government continues to consult on its regulatory response to generative AI tools that have become pervasive across the economy in less than a year.
Jointly led by the Digital Transformation Agency and the Department of Industry, Science and Resources, the new AI use in government taskforce will seek to ensure agencies can use AI safely and responsibly.
Since the arrival of OpenAI’s ChatGPT in November last year, agencies have largely been left to their own devices when it comes to AI governance, with some agencies – like Home Affairs – recently opting to block the generative AI tools.
In July, the DTA released advice to guide the public sector on the use of generative AI, which chief executive Chris Fechner last week said was intended to prevent agencies from “bleed[ing] out” while the government’s thinking matures.
The advice permits agencies to experiment with generative AI tools “in cases where the risk of negative impact is low”, but warns against using generative AI tools for service delivery and decision-making.
But the advice is only voluntary and overridden by internal IT policies, effectively making agencies the “arbiters of what is acceptable”, according to the Australian Information Industry Association.
The new taskforce will work create the necessary guardrails for AI over the next six months, as well as supporting advice, direction, and tools in order for agencies to leverage the emerging technology safely.
Technical and policy experts from the Australian Taxation Office, Department of Finance, Department of Health and Aged Care, Department of Agriculture, Fisheries and Forestry and Department of Home Affairs will contribute to the multi-disciplinary team.
The Tax office is one agency already using machine learning to identify fraud and superannuation underpayments using its internal dataset, but is also looking ahead to the role that generative AI will play.
Finance minister Katy Gallagher described the taskforce as an “important step” to make sure the public service takes advantage of the technology, while addressing concerns over its use.
“AI has the potential to improve productivity within the APS and make government services better for all Australians,” she said.
“However, as the Robodebt Royal Commission has shown, the unmitigated use of advanced technologies also has the potential to cause harm and undermine trust in government.
“This Taskforce will investigate a policy for the safe and responsible use of AI in government to ensure we can reap the rewards of AI while safeguarding against potential risks.”
Ms Gallagher has previously spoken about the importance of the governance and ethical implications of an AI system, particularly “where discretion is used to determine an outcome”.
“In many decisions, human insight and intelligence as well as knowledge of context and history is a crucial part of decision-making and should not be outsourced to AI,” she told the Impact 2023 conference in April.
Industry and Science minister Ed Husic, who’s department is leading the government’s economy-wide consultation on safe and responsible AI, said the taskforce “combines a wealth of experience in the field of AI with broader perspectives from across the APS”.
“We’re witnessing the capabilities of generative AI model evolving by leaps and bounds every few months,” he said.
“It’s our job to make sure that we don’t just get the maximum benefit from the tech but have solid plans in place to protect our government systems from emerging risks now and into the future.”
Do you know more? Contact James Riley via Email.
The idea is sound, but filling the task force from people within the Public Sector only is yet another potential case where the best advice is not sought. Where is the industry leading AI Ethics leadership going to come from?