The Federal Court of Australia is currently considering the patentability of computerimplemented inventions in the cases of Research Affiliates and RPL Central. Research Affiliates concerns patentability of a method, system and computer program product for passive investing.

The physical result of the method is a file containing a financial index. RPL Central concerns a method for assessing competency or qualifications of individuals against recognized standards (‘Recognition of Prior Learning’ or ‘RPL’).

The Research Affiliates claims were held unpatentable in both the Patent Office and Federal Court on appeal. On the other hand, the Federal Court has decided that the RPL Central claims held unpatentable by the Patent Office are patentable. What contrasts the two decisions and what lessons do they provide for drafting patent specifications for computer-implemented inventions in Australia and more broadly, the management of this type of intellectual asset in Australia?

An artificial state of affairs is important

The Research Affiliates refusal partly resulted from the Applicant’s admission that the physical result generated by the claimed method was a file containing a financial index, i.e. data alone. Untransformed data, of itself, cannot create the ‘artificial state of affairs’ necessary for patentability under Australian law. This finding was bolstered by insufficient description of how the computer implemented the investment method to create the ‘artificial state of affairs’ or ‘physical effect’ necessary for patentability. Whether or not the investment scheme presented a ‘paradigm of economic endeavour’ was considered irrelevant in this case.

The RPL Central case provided a significant factual contrast. The Court remarked, at some length, on the detailed description of improvements in existing RPL processes which involved the automated generation of a wizard or similar user interface to perform administrative work necessary to gather evidence from a prospective candidate and enable qualifications recognition from a wide range of training organisations. The Court found a useful result in the capacity of the invention to overcome difficulties in seeking out relevant education providers and the necessary information to enable recognition of prior learning. The invention provided the necessary ‘artificial state of affairs’ because data was not just collated or indexed, but actually transformed. The Court also found that the invention provided economic benefit to the educational sector of the economy. The invention was found patentable.

How to draft computer-implemented inventions in future

Some practitioners in the computer-implemented invention field have found it difficult to reconcile the two decisions, which are now both under further appeal. At the time of writing the Australian Patent Office is stubbornly applying the Research Affiliates approach. Subjectivity is perhaps unavoidable in this area as minds differ as to where the line between patentable and unpatentable invention lies. Nevertheless, this writer sees some useful guidance on drafting patent specifications for computer implemented inventions in these Federal Court decisions.

At least until the full bench of the Federal Court decides otherwise, computer-implemented inventions should be considered patentable if the learnings from RPL Central are adopted. The specification must describe the computer-generated processes in implementing the invention. Where a method with a number of steps is involved, each step should require or involve a computer generated process. The specification should clearly demonstrate the physical effects of the invention in the form of a change of state or information in part of the computer (at least). There are echoes here of the US originating machine or transformation patentability test.

In looking so closely at the specification to assess patentability, well established principles that the invention must be defined by the claims are seemingly altered. The Australian Patent Office and Courts are looking to more than the claims and to description of the invention to clearly demonstrate working of a computer as a machine to transform information into an economically useful ‘physical effect’ or ‘artificial state of affairs’. Patent specifications need to be drafted to meet these requirements.