Govt issues guide on R&D Tax Incentive eligibility for AI-related work

The Department of Industry, Science and Resources has issued a guide to help Australian businesses understand if the AI work they’re doing qualifies for the R&D Tax Incentive (R&DTI).

The guidance explains when activities like developing, training and testing AI models may be eligible for the R&DTI.

The guide said AI R&D activities may involve scoping and designing AI-enabled solutions, developing and training models, testing and evaluating model performance, integrating AI into software applications and platforms or deploying models into operational use.

The Department also said there are some common risk areas organisations should be aware of when self-assessing eligibility for the R&DTI. For an AI-related activity to be eligible as a core R&D activity, it must aim to resolve a technical hurdle where a proposed solution (the hypothesis) is tested through planned experiments.

It said while many activities may be complex, they may not meet eligibility criteria because they can generally be addressed by experts in the field using existing knowledge and solutions.

Activities that are unlikely to qualify include implementing logging, alerts, dashboards, or routine performance checks using established methods and tools to confirm expected model behaviour; cleaning, formatting, and aligning data to meet a model’s documented input requirements using standard data preparation techniques, where the transformations required are known in advance; or running regression, acceptance, or functionality tests using established testing methods to confirm a system works as intended, where expected outcomes are already known.

In addition, adjusting or fine-tuning parameters using established methods when it is known that the issue can be solved by parameter changes and the effect of changing the parameter is well understood; as well as integrating known model outputs (such as scores, labels and recommendations) into applications or dashboards, using pre‑defined logic, rules and interfaces, where the behaviour of the model and the integration approach are established, are also unlikely to meet the requirements of core R&D activities.

The Department said AI‑related activities are eligible R&D activities when they meet the requirements for core R&D activities or supporting R&D activities.

These requirements do not depend on the technology used, however, therefore using AI in software development does not make an activity eligible.

AI-related activities may meet the requirements of a core R&D activity where a technical hurdle exists and an expert in the field considers that only experimentation will determine if a proposed solution, or the way to develop a solution, can resolve the technical hurdle.

The Guide also provides several examples to assist organisations understand their eligibility for the R&DTI if they’re doing AI-related work.

The Department said organisations wanting to understand where AI fits into the R&DTI should also review the Software development sector guide for the R&DTI page.