The Australian Government will appeal a Federal Court decision to allow machines to be recognised as inventors on a patent application, arguing the recognition is “incompatible” with Australian patent law.
Dr Stephen Thaler, a US-based developer who is part of a global project to have artificial intelligence (AI) recognised as inventors, had his patent application rejected in Australia in February because the documents failed to name a human inventor.
He sought a judicial review of the decision and in July the Federal Court ruled in his favour, saying too restrictive a view of the term inventor would “inhibit innovation”.
IP Australia has confirmed it will appeal the decision, which was based on an interpretation of the Patents Act 1990. But the agency said the move does not represent a government policy decision on whether or not AI should be considered an inventor of a patent.
“The appeal is centred on questions of law and the Commissioner’s interpretation of the patents legislation as it currently stands,” a spokesperson for IP Australia told InnovationAus.
The Australian legislation does not define the term “inventor”, meaning it is considered to have its ordinary meaning.
The Australian Commissioner of Patents has argued this meaning is inherently human, and the reason Dr Thaler’s application was originally rejected.
In July, Federal Court Justice Beach said the Commissioner “is incorrect in saying that you cannot have a non-human inventor” and found AI can be named as an inventor. However, he also found an AI system cannot apply or receive a patent.
“In summary, in my view, an inventor as recognised under the Act can be an artificial intelligence system or device. But such a non-human inventor can neither be an applicant for a patent nor a grantee of a patent,” Justice Beach said.
Dr Thaler is a developer of Device for the Autonomous Bootstrapping of Unified Sentience (DABUS) and part of a global Artificial Inventor Project seeking intellectual property rights for inventions generated by AI.
His Federal Court victory in Australia marked the first decision by a court anywhere to allow a patent naming a robot inventor to proceed.
Dr Thaler was represented by global law firm Allens in Australia, with local partners Lauren John and Richard Hamer running the case.
Mr Hamer said the case is important because AI is a “fundamental part of the fourth industrial revolution”.
“Patents aim to encourage innovation and innovation by AI systems has the potential to be at least as valuable as innovation by human inventors. It should be encouraged for the same reasons.” he told InnovationAus.
“It is to be hoped the decision will be an example in other countries.”
Similar applications have been filed in Canada, China, Europe, India, Israel, Japan, South Africa, the UK and the US, as part of the Artificial Inventor Project.
In South Africa, AI has been recognised as a legitimate inventor, but at the administrative level rather than in court, and the patent could still be revoked for a number of reasons.
A US judge last week ruled AI could not appear as inventors on US patents, upholding the country’s patent office decision.
With different rulings in different jurisdictions, Mr Hamer said some patent filers which have relied on machines for their invention may seek to find other humans to be named as inventors.
“That risks the validity of the patent if it is untrue,” he said.
“In the long term, the aim would be international alignment. This decision [in Australia] is only the very start of that process.”
The Federal Court decision was “a strong one by a well-regarded patent judge”, Mr Hamer said. “We hope the decision will be upheld.”
IP Australia declined to comment further on the case because it didn’t wish to pre-empt the court’s decision.
“The decision to appeal does not represent a policy position by the Australian Government on whether AI should or could ever be considered an inventor of a patent,” an IP Australia spokesperson said.
Do you know more? Contact James Riley via Email.
It seems some members of the judiciary wish to confirm the old aphorism: ‘The Law is an ass’.
The proponents of the status of “AI” as an inventor need to answer:
– what do they mean by “AI”. By what criterion and in what way is intelligence being claimed?
– if they mean software, are they claiming this for all versions of the software, or only some specific version or versions? Is it still being claimed if the software is later found to contain ‘bugs’ or errors?
– if the same algorithm(s) is/are implemented in different languages, are all implementations AI? Are they ‘the same’ AI, even if written in different languages?
– what if the algorithm is changed, e.g. changing among for loops, while loops and do loops? I.e. it’s rewritten to have the same function(s) using a different implementation. Is this still the same AI? Or is the algorithm abstracted from its implementations the AI. In which case it can’t do anything, since it’s an abstraction, not executable code.
– are they referring to the source code? Or to compiled interpreted code, or compiled machine code? If so, then none of these actually do anything until they are executed. So they are not inherently intelligent, since they are inherently inactive.
– are they referred to executing code? If the code is executed more than once, does this constitute more than one AI?
– If the code is executed on more than one machine, does this constitute more than one AI?
– or is it the machine executing the code that is artificially intelligent? What then of the same code executed on different machines?
– or is it in any case not the supposed AI that created the invention, but the person(s) who invented the AI?
– and how is this different from the use of machines and equipment to create an invention, e.g. a processing lab that mechanistically combines a series of proportions of ingredients to produce one combination that has the qualities of interest. Is the machinery then the inventor?