Patent battles

What counts as an original invention? What does it mean to be an inventor? If a data-crunching, machine-learning artificial intelligence system generates something that its human programmer never conceived, can that invention be patented?

These are just some of the thorny questions patent law is currently grappling with, only a few short years after the dust had seemingly settled on a decades-long debate that ultimately found that genetic material, being a product of nature, was not patentable after all.

Meanwhile, bitter patent wars rage over who invented revolutionary editing technologies, and a legal stoush looms over who owns the rights to manufacture COVID-19 vaccines that were developed, remarkably, in less than a year.

Patent disputes around emerging and disruptive technologies, from sewing machines to heart stents, have a long, abrasive history. But never before has patent law seen something quite like the challenge posed by artificial intelligence (AI) systems, which mimic human brain networks often with surprising or insightful results.

“The law really struggles in understanding how to deal with these highly abstract concepts that are outside of normal human experience,” says Justin Blows, an optical physicist turned patent attorney who specialises in commercialising computer technologies.

“The legal system never imagined an inventor that wasn’t human. Who would have thought that 200 years ago?”

Borne out of the steam-powered Industrial Revolution, patents are designed to protect the intellectual property of (human) inventors and to incentivise innovation. Owners are granted exclusive rights to their invention for 20 years; in return, competitors, and by extension society, gain access to that technological know-how for a fee.  

Fast-forward nearly two centuries, and the law is trying to catch up to science as it pushes the boundaries of what’s possible in the realm of human-like machines and quantum entanglement. The same basic principles of patentability apply – to patent a piece of technology it must be new, useful and involve an unforeseen inventive step – but those principles are being tested in myriad ways.

“There’s been a lot of issues with trying to adapt a patent regime designed in the industrial age for all the new developments that have been taking place,” says Matthew Rimmer, a professor of intellectual property and innovation law at Queensland University of Technology.

“Often there is very little guidance in what has gone before in the law,” adds Blows. “Judges as a result struggle because the whole system is based on precedent.”

The limits of patent law are being tested. In 2018, American computer scientist Stephen Thaler quietly began filing patents for a few handy inventions – a safety beacon and an extra-grippy container – conceived by an artificial intelligence system he developed. Provocatively, Thaler listed the AI system, DABUS, as the inventor in his patent applications, not himself.

Then, in a judicial world first, a judge in Australia’s Federal Court ruled in Thaler’s favour, finding that an AI system, a machine, could be an inventor in its own right. Less than a year later, however, the controversial decision has been overturned, putting Australia back in line with the US, New Zealand and European patent offices that rejected Thaler’s claims, arguing an inventor must be human.

“There are certainly real tensions in terms of the way in which the legal system deals with new breakthrough technologies,” says Rimmer. “But there are also different philosophical and ideological approaches with how judges deal with some of these matters.

“As a student back in the 1990s, we had very similar debates about how the patent system would deal with computer programs, information technologies and this newfangled thing called the internet.”

Software was controversially deemed unpatentable because it was more like mathematics than an invention; this has created a black spot, in that the code itself of AI systems cannot be patent-protected. But Blows notes that the law is changing, albeit slowly: “What comes out of courts are opinions, not final decisions.”

Natalie Stoianoff, a professor of intellectual property at the University of Technology Sydney, says the law is operating as best it can, with courts nutting out how to apply the principles of patent law, case by case.

“There’s always going to be a period of time where the law has to adjust and consider how to apply the principles [of patent law] to different technologies,” says Stoianoff. “It’s a matter of the legislation being interpreted by judges in cases to further develop the way the law actually operates.”

But as a result, the law lags far behind science, with patent disputes over new technologies taking years or even decades to work their way through patent offices, trial cases, appeal hearings and up to superior courts.

All sorts of arguments are taking place around the world about whether AI systems can “invent”, Stoianoff says. “Ultimately, there will probably need to be some sort of international agreement on this.”

Complicating matters is the fact that artificial intelligence and CRISPR gene-editing technologies are still in their infancy but spawning all manner of applications. “Both fields have a lot of hype around them, and sometimes it takes a long time for the mature applications to appear,” says Rimmer.

More than 11,000 families of patents on CRISPR-related gene editing technologies have now been filed. Meanwhile, the bitter patent brawl rages on between two rival teams who both claim they invented CRISPR first – with no end in sight.

Companies vying to commercialise their applications of CRISPR technologies could be hamstrung by the unfolding patchwork of patent law rulings, some experts warn. The legal hubbub “casts a shadow over whether and when the first CRISPR therapies will be launched onto the market,” University of Illinois law professor Jacob Sherkow told The Boston Globe.

But others argue that the COVID-19 crisis has shown that patent rows, although costly and convoluted, have not kept therapies from reaching the market in the grips of a pandemic. Who can afford them is another matter.

Another fundamental debate bubbling up is how well the patent system is working to incentivise research, or if the law is actually an obstacle to scientific progress. Without patents to protect their technologies, no start-ups would attract funding to commercialise research. But are patents a barrier to accessing knowledge and technology?

“Often you have to look at what’s happening in particular sectors,” says Rimmer. “There has always been a lot of debate about the purposes of patent law and what objectives it is meant to serve.”

Some critics say patent protection stifles scientific progress by creating monopolies that deny access to technologies or restrict knowledge sharing which would otherwise spur competition and discovery. Others see it differently: “Today’s patent is next year’s footnote, not fetters permanently restricting research and public benefit,” as University of Canberra law professor Bruce Baer Arnold put it.

But debates about who ultimately benefits from patents have intensified in recent years, Rimmer says, with the pharmaceutical industry lobbying for law reform to lengthen and strengthen patents to protect their intellectual property rights. This has incited backlash from others who argue that public interests – in the form of access to affordable medicines – should be better respected.

“There have been some real clashes over what should be protected under patent law and what should be left in the public domain, or at the very least open to open licensing,” Rimmer says. One of the most recent manifestations has been the high-stakes dispute over who is the rightful owner of Moderna’s COVID-19 mRNA vaccine, when the US National Institute of Health funded the pharmaceutical newcomer’s early research.

What’s troubling to legal experts like Rimmer is that debates about waiving patent protection to ensure the equitable access to essential medicines were foretold in the HIV/AIDS crisis, revisited when tuberculosis, malaria and Ebola outbreaks hit, and yet the same issues flared up, unresolved, in the COVID-19 pandemic.

“It’s startling to me that even a major global public health and economic crisis has not been enough to overcome certain inertias within the patent regime,” Rimmer says.

Then there is a whole other curly question about whether the patent system is capable of making good judgements about the relative contribution of collaborating scientists involved in huge international, discovery-making teams, which have become the norm in many fields.

“In the age of big science projects, we tend to be a world away from patent law focusing on individual lone inventors making machines and contending with very complex collaborations and rivalries,” says Rimmer.

“We need to work out a way for the patent system to be much better at doing that.”