Home » News » Assessment of AI related activities under the R&D Tax Incentive and potentially emerging compliance focuses

Assessment of AI related activities under the R&D Tax Incentive and potentially emerging compliance focuses

Artificial Intelligence (AI) has been a hot topic in recent years with more companies seeking to utilise the technology in order to capitalise on gains in capability and efficiency.

Increasing numbers of companies are investing in AI developments and anecdotally, this probably means an increasing number of companies submitting claims for AI related developments under the R&D Tax Incentive.

However, whilst AI is a relatively new and developing field, companies need to be careful not to assume that projects involving AI will automatically qualify under the R&D Tax Incentive (in the same way that companies undertaking software development activities are not necessarily automatically eligible).

Eligibility of companies’ R&D activities will come down to assessment of the activities and their underlying supporting documentation against the law.

We expect that AusIndustry may release more dedicated guidance on the assessment of AI R&D activities in time, however in the interim; the principles in the past and current software guidance are relevant. Notable extracts from AusIndustry software guidance over the years includes:

• Eligible core R&D is not learning how to use existing products, technologies or techniques in the manner in which they are designed to be used. Eligible R&D is not using such products, technologies or techniques in a specific application;
• Software development activities can pose a challenge to the self-assessment of eligible activities because the process of developing, modifying or customising software is superficially similar to the eligibility requirements for core R&D activities under the R&D Tax Incentive. They are by definition systematic and can be iterative and cyclical, and almost always involve testing. However, they are not necessarily experimental as required under the programme’s legislation;
• The identification of technical uncertainty in a project does not automatically identify a need for R&D. In many cases technical problems are resolved by applying the expertise and knowledge of the development staff or by selecting products that are available to solve the problem;
• When assessing whether there is an unknown outcome, you need to consider:

o    whether an outcome is scientifically or technologically possible, or how it can be determined

o    whether existing knowledge or capability can be adapted to solve a problem

AusIndustry software guidance also now includes a HypotheticalMachine Learning case study which may be useful.

Where we see scope for potential pitfalls in AI claims is where companies may be applying known processes (including established Software development, AI, ML and data science processes) to their commercial parameters and workflows, but may be doing so in a standard or known manner. For example, a company that customises or adapts existing or known AI technology to automate their business operations without generating any new technical knowledge or capabilities would be unlikely to be conducting R&D activity. AusIndustry may argue in this instance that:

• the company is not generating new knowledge in the form of a new product or process, rather, they are applying existing knowledge to their business operations;
• there is no scientific unknown outcome being investigated via experiments;
• the testing or language model training activities the company is conducting may relate to resolution of business challenges rather than generation of new knowledge.

Current compliance reviews indicate an increasing focus of AusIndustry in seeking to understand the processes companies go through BEFORE commencing their experiments in order to justify:

• That they are seeking to generate new technical knowledge;
• That there was an unknown technical outcome which could not have been determined based on existing knowledge;
• That the company made effort to search for and document existing knowledge, then to analyse whether it could have been applied to avoid the need for them to conduct experiments;

The outcome of compliance processes may depend on the quality of explanation a company can provide around the new knowledge and unknown outcomes, supported by evidence of an attempt to search for existing knowledge, such as:

• Correspondence with professionals or experts (emails, reports, call notes);
• Internet/google searches;
• Review of scientific, technical or professional literature;
• Technology reviews;
• Patent and other searches;
• Screenshots of questions posted on tech blogs or forums;
• Details of failed attempts to use existing technology.

AI being a relatively new and developing field may impact the level of existing knowledge that is available to companies seeking to conduct experiments, however companies must still evidence an effort to search for existing knowledge at the outset of their R&D activity to be able to explain how their activities will generate new knowledge.

Companies should also keep an eye on the AusIndustry Software Guidance for any developments.

Please get in touch with our office if you require assistance, would like to speak to someone about a potential claim, or check out our website for more information.