Rokt decision boosts certainty on CIIs in Australia

In-house counsel at artificial intelligence (AI), medical device, semiconductor and entertainment companies say that a Full Federal Court of Australia ruling has brought much-needed clarity on patentability of computer-implemented (CII) inventions.

In a May 21 decision, the court in Commissioner of Patents v Rokt overturned a Federal Court of Australia decision in which Rokt's digital advertising platform was found to be patentable subject matter.

James Wan, general counsel at an AI company in Sydney but who is expressing his personal views, says that the Rokt decision at the Full Federal Court wasn't surprising given the recent Encompass Corporation v Infotrack ruling. In that 2019 decision, the Federal Court confirmed that computer implementation is insufficient in providing patentability to an unpatentable scheme.

"It was actually surprising for Rokt to have won at the Federal Court," says Wan, adding: "These two decisions are consistent and provide clarity on patentability of CIIs in Australia."

He says that the leading case, Commissioner of Patents v RPL Central, made it very clear back in 2015 that there needs to be a technical improvement made in the computer technology for an invention to be patentable. It cannot just be a scheme requiring computer implementation.

The Encompass and Rokt decisions, within a one-year period, maintain this position. These developments are also consistent with those in other countries, such as the US.

"For patent applicants, it would be best to keep in mind that schemes by themselves are not patentable," Wan says. "They have to solve technical problems, rather than [be] just a business idea that provides commercial advantage."

The in-house counsel at an Australian entertainment company says that with the Rokt decision, the court has given practitioners a litmus test which can be used to quickly reject CII patentability.

"It gives a glimmer of hope on how one can fall out of the grey area of CIIs," he says. "It means that applicants have to go the extra step of describing and even claiming any particular programming tricks if the invention is software heavy, or the essential components of the invention are hardware heavy."

The counsel continues: "This avoids any arguments that either or both the software and hardware are mere vehicles to implement an idea. Many have known that this extra step can be the saving grace, though what is required to get out of this mere vehicle is still anyone's guess."

He says that if he needed to sum it up, software-aided business methods are out, but software-controlled dedicated hardware – if properly characterised – hopefully remains in.

Pasquale Aliberti, head of patents at semiconductor startup Morse Micro in Sydney, says that the Rokt decision clarifies that patentability should be assessed based on a sound construction of the claims and that neither the prior art nor the common general knowledge should be imported into the patentability assessment.

"The decision does not rule out all CIIs from being patentable," he says. For example, CIIs directed to new technical methodologies or those defining new systems that can be used to solve a technical problem may still be patentable in Australia.

Cecillia Suatan, senior legal counsel, IP and commercialisation at a medical device company in Sydney but who is expressing her personal views, says that the Rokt decision provides a clear direction on the patentability of CIIs in Australia.

"The test of whether the invention merely uses generic computer functionality to implement an abstract idea is relatively straightforward compared to, for example, the two-part Alice test in the US, which could lead to inconsistent results," says Suatan.

She adds that in business, predictability of the outcome – even if it's adverse – is valuable because it takes the guesswork out of decision making.

"Businesses that have been routinely filing CII-related patents may be far less inclined to file in Australia, and may choose to re-direct their patenting efforts to more favourable jurisdictions such as China, Japan and South Korea," says Suatan.

She continues: "If they choose to continue to file in Australia, they may need to devote more resources upfront to careful claim drafting that does not fall foul of Rokt and its predecessor decisions."

She says it is important that the claims are directed to improvements in computerisation, such as in its programming or operation, rather than recite features and functionalities that are common to all computers.

Suatan says that there will be more of an emphasis on keeping CIIs protected as trade secrets. But software – both its code and graphical user interfaces (GUIs) – continues to be protectable by copyright. She says that GUIs may also be protected as a registered design in Australia, but the 10-year limit on protection may be a disincentive to file.

Duncan Longstaff, principal at Shelston IP in Sydney, says that patentees should focus on understanding and describing the detail of how a CII works rather than solely on what it achieves or the end result.

"Engaging with the inventors and programmers or technical staff when having the application drafted is good practice as it provides good technical description as well as insight into the unique details of the inventiveness," he says.

This information can then be used to support the patentee's arguments and assertions about the technical effects and advantages when prosecuting the applications. However, he says that this may not be a straightforward process and that patent drafting and prosecution in accordance with these principles will often still prove difficult.

"It may be significant to emphasise the unique technical operation of the invention, and to de-emphasise any business application it may have," he says. "Also, rather than saying an application can also be implemented on a generic computer, it may be best to provide some guidance on possible alternative implementations on alternative computer architectures."

Longstaff says that it is necessary to avoid using only functional language to describe the result or purpose of a step in an invention. An application should describe how the invention is actually performed and implemented, in a technical sense, rather than broadly describing an idea or information flow. Specifications should include sufficient technical detail on how any hardware, computer functions or software used for carrying out the invention actually perform their technical functions and improve computer technology.

As a practical tip, he says that describing the details of a particular algorithm or technical effect may reduce the risk of the invention being characterised as abstract or requiring mere generic computer implementation. Similarly, figures should have detail and support the technical description of the invention.

The Full Federal Court of Australia's decision in Rokt has given much-needed clarity on patentability of CIIs. Australian patent applicants would do well to focus on the improvements in computerisation and the technical improvement of an innovation, rather than on the computer scheme itself.