You have 0 free articles left this month.
Register for a free account to access unlimited free content.
Powered by MOMENTUM MEDIA
lawyers weekly logo
Advertisement

ATO weighs up adopting AI to assist 800 core developers with programming

Technology

The Tax Office wants to free its developers from trivial tasks and allow them to focus on work that is “higher value”.

By Daniel Croft 7 minute read

The ATO is weighing up adopting AI for its core developers, allowing them to use the technology to speed up their workflows and help them eliminate common programming roadblocks.

The ATO said the AI coding tool it’s looking at will free up its 800 core developers from trivial tasks and will assist with bug fixes, code suggestions, codebase refactors across platforms and automated test and script generation, allowing them to “focus on higher value work”, like application security controls, maintenance of existing technologies and test case planning, according to an ATO request for tender as seen by ITnews.

“The ATO has a complex, multi-technology IT delivery environment, supporting over 800 core developers, spanning multiple technologies,” the agency said in its statement of requirements.

The agency said the AI it uses must integrate with the ATO’s Azure DevOps and Git repositories and be able to translate legacy code into modern languages.

It also needs to integrate with Microsoft development environments such as Visual Studio Code and Visual Studio 2019 and 2022.

Most notably, any code processed by the AI cannot be used to train the model further, with the ATO citing privacy and compliance issues.

Ying Yang, ATO’s assistant commissioner for data science, said the use of large multimodal AI models to assist in auditing taxpayer-submitted documents is already being experimented with.

 
 

“Why large multimodal instead of large language? The documents are not always in plain language,” she said at the AI Innovation Showcase in Canberra.

“It can be images as well, so the team is right now playing with large multimodal models to see whether it will increase the performance of the document understanding.”

You need to be a member to post comments. Become a member for free today!
You are not authorised to post comments.

Comments will undergo moderation before they get published.