TLDR: The Australian Taxation Office (ATO) is actively assessing artificial intelligence (AI) coding assistant tools to enhance the productivity of its 800 core developers. This initiative aims to automate routine programming tasks, allowing developers to focus on higher-value work like security and strategic planning. The ATO is seeking a secure, integrated SaaS solution that will not store or use its code for AI training, aligning with its broader strategy to embed AI across its operations.
The Australian Taxation Office (ATO) is embarking on a significant technological upgrade, exploring the integration of artificial intelligence (AI) coding assistant tools to bolster the efficiency of its approximately 800 core developers. This strategic move, highlighted in recent requests for tender, aims to streamline programming challenges and enable the ATO’s development team to concentrate on more complex and critical tasks.
The envisioned AI coding assistant is expected to offer a suite of advanced functionalities. These include intelligent code suggestions, automated bug remediation, the generation of test cases and scripts, and the capability to refactor codebases across diverse technological platforms. A key objective is to free up developers from repetitive coding tasks, allowing them to dedicate more time to high-value work such as test case strategizing, application security protocols, and the maintenance of existing technologies. The ATO noted in its requirements statement that it ‘operates within a complex, multi-faceted IT delivery landscape that supports over 800 core developers across a variety of technologies.’
Technical specifications for the desired solution are stringent. The ATO is seeking a software-as-a-service (SaaS) solution that must seamlessly integrate with Microsoft’s development environments, specifically Visual Studio 2019 and 2022, along with the Visual Studio Code editor. Furthermore, the tool must connect with the agency’s Azure DevOps and Git repositories and possess the ability to translate legacy code, such as COBOL, into contemporary programming languages.
Security and data privacy are paramount concerns for the ATO. The tender explicitly stipulates that any code processed by the AI tool will not be stored or utilized for training purposes, directly addressing potential privacy and compliance issues. The solution must also incorporate ‘guardrails’ to mitigate the risk of harmful material or malicious code suggestions being input or output from foundation models. Additionally, the system must be hosted in a secure ATO-approved environment (on-premises, Azure, or AWS) within an Australian region, ensuring data encryption both at rest and in transit to comply with ATO’s data privacy policies and Australian Public Service (APS) legislative requirements.
Also Read:
- Australian Workers Secretly Embrace AI Amidst Policy Gaps, New Report Reveals ‘Shadow AI’ Risks and Benefits
- New Report Reveals GitHub Copilot Not the Foremost AI Tool for Developers
This initiative is part of the ATO’s broader strategic push to embed artificial intelligence throughout its operational framework. Beyond developer assistance, the agency has identified enterprise-level AI use cases including fraud detection, client risk profiling, and document understanding. ATO Assistant Commissioner for Data Science, Ying Yang, has previously detailed the agency’s trials of large multimodal AI models to assist in auditing taxpayer-submitted documents, underscoring the ATO’s commitment to leveraging AI for enhanced efficiency and compliance across its diverse functions.
