The Digital Transformation Agency (DTA) has announced its plans to explore how artificial intelligence (AI) can support panel evaluations, particularly in high-volume rounds.
The agency said that over the past 18 months, it has conducted "extensive" research and testing, including collaboration with CSIRO’s Data61 unit, alignment with AI technical standards and the DTA AI policy team, and internal trials and assurance activities.
The trials revealed that AI can help streamline evaluations, especially when there are large numbers of applications.
"Importantly, AI is used to assist, not replace, human evaluators," the DTA said.
The AI will not mark applications as unsuccessful or determine value for money.
Every AI-generated score will be paired with a human evaluator, and if the AI and human scores are closely aligned, the AI’s assessment supports the decision.
If there’s a significant difference or the case study is rated unacceptable, the AI’s score is set aside and a second human evaluator reviews the application.
If agreement still can’t be reached, the Evaluation Chair makes the final determination. This ensures that every decision is made with experienced human oversight and contextual understanding.
AI contract template released
Unrelated to the DTA's own use of AI, the DTA also released an AI contract template to support agencies in procuring AI systems and services.
The template builds on the AI model clauses released earlier in the year and offers a starting point for managing AI-specific risks in contracts.
The document includes guidance on clause selection for different use cases, explains the rationale behind optional clauses, and provides prompts to assist agencies in applying them effectively.
The template will be updated over time as AI technology and policy landscapes evolve and currently covers commercial-off-the-shelf AI systems delivered as cloud services.
_(11).jpg&h=142&w=230&c=1&s=1)
.jpg&w=100&c=1&s=0)
_(8).jpg&w=100&c=1&s=0)
.jpg&q=95&h=298&w=480&c=1&s=1)
.jpg&q=95&h=298&w=480&c=1&s=1)
