A Guide to the R&D Tax Incentive for AI Founders
The rapid proliferation of AI has led to a surge in R&D Tax Incentive claims for AI and ML-related activities. To ensure companies are compliant, it’s crucial to understand how these activities are treated under the R&DTI, when they qualify for registration.
Benefits of Claiming the R&D Tax Incentive
The R&D Tax Incentive offers a refundable tax offset of up to 43.5% for eligible R&D activities, for company groups with aggregated turnover less than $20m. For AI and machine learning, which often require significant computing resources and expensive hardware, this refund can significantly improve cash flow for start-ups. Eligible expenses may include:
- Engaging experienced data scientists and engineers including employees or contractors
- Working with Research Service Providers
- Depreciation for computer or other hardware used in devlopment
- Supporting infrastructure, such as cloud computing services
To claim these costs, businesses must maintain evidence demonstrating a clear link (nexus) between the expenses and the claimed activities.
Classification of AI/ML Activities
The R&D Tax Incentive application requires businesses to specify the relevant Australian and New Zealand Standard Research Classification (ANZSRC) code. While many AI/ML founders might assume “Artificial Intelligence” or “Machine Learning” (4611) is the appropriate code, its scope is limited. These designations cover the development of underlying techniques used in machine learning. Machine learning or AI applied to a specific domain, is more correctly classified under the corresponding domain’s code.
For exmaple:
- Projects developing underlying machine learning techniques: Classified as Machine Learning research.
- Projects applying machine learning to a specific domain (e.g. medical diagnosis using biomarkers): Classified as research within that domain (e.g. Biomedical and Clinical Sciences).
- Projects involving the development of complex computer systems or architecture involving machine learning or AI: May be included in other Information and Computing Sciences domains such as Software Engineering.
This distinction is critical because the R&D Tax Incentive excludes certain research areas, including social sciences, arts, humanities, and market research into consumer behaviour and preferences. Consequently, activities applying machine learning to answer research questions in these areas may be ineligible for registration. The inclusion of AI or ML techniques does not, in and of itself, constitute eligibility for the R&D Tax Incentive.
The requirements for activities to be conducted for the purpose of generating new knowledge, and to have uncertain scientific outcomes, should be referenced to the specific field of inquiry. For example, developing an ML model to determine consumer preferences may be excluded.
Unknown Outcomes in a Rapidly Evolving Field
The R&D Tax Incentive requires applicants to demonstrate that the activity’s outcome was unknown and could not be determined in advance based on existing knowledge, information, and experience. Given the rapid advancement of machine learning, this can be challenging for AI founders.
The “Hypothetical Machine Learning case study” published by AusIndustry suggests a suitable process to determine the lack of available knowledge:
- Preliminary Literature Review: Review research articles in the field to identify unanswered research questions.
- Hypothesis Development: Formulate an initial hypothesis based on the identified gaps.
- Secondary Literature Review: Conduct a second review to find any follow-up research testing the proposed solution or hypothesis.
If the second review finds no research testing the solution, the business can conclude that the activity’s outcome was unknown based on current knowledge.
Record Keeping
Businesses must document their research activities, including:
- Documented discussions explaining the rationale for undertaking the activities.
- Results from online research, including article links and notes/discussions about the information found.
- Plans outlining how experiments will be conducted, including the hypothesis.
- Results of all experimental runs and analyses of the results.
- Timesheets or other verifiable time allocation documents should also be kept where employees or contractors are involved in R&D and non-R&D activities.
By understanding these guidelines and maintaining thorough records, AI founders can ensure their compliance with the R&D Tax Incentive and maximise the benefits of registering their activities.
For support with R&D tax incentives and grants, please reach out to our dedicated team.
