In early October DeepMind, a British-American artificial intelligence research laboratory which serves as a subsidiary of Google, and German drug maker BioNTech separately announced that they will be launching AI lab assistants to help advance science. At the same time BioNTech’s London-based AI subsidiary InstaDeep presented AI models that could help the drug maker identify or discover new targets to tackle cancers.
The announcements underscore how AI could accelerate scientific discoveries but also raises a host of thorny questions: If human scientists collaborating with AI lab assistants makes an important discovery – say a drug that can cure a lethal disease – who should receive the patent: the person or company who developed the AI assistant, the scientist who instructed and collaborated with the AI assistant, the AI assitant itself or a combination of all three? And if such a discovery can be made in a day or even minutes, at a fraction of the current cost, do we need to rethink the whole economic model underlying drug discovery and prescription drug pricing and ensure that the patent system grants early and fair access to people around the globe?
As AI accelerates scientific progress and patent creation, the clash between the human right to benefit from scientific progress and its application (the human right to science) and the right for innovators to protect their inventions is intensifying. How can we design future protection models that balance innovation, access, and fairness in a rapidly evolving technological landscape? That is the main question that will be posed during a panel discussion which will be moderated by The Innovator’s Editor-in-Chief at the invite-only annual summit of The Geneva Science and Diplomacy Anticipator (GESDA), on Oct. 9-11.
Under the theme of “The Great Scientific Acceleration” GESDA, a Swiss Foundation initiated by the Swiss Federal Government and the City and Canton of Geneva, whose vision is to “use the future to build the present” by bringing together different communities to jointly anticipate scientific and technological advancements and, based on them, develop inclusive and global solutions for a sustainable future, will convene 1,000 scientists, diplomats, policymakers, innovators, executives and citizens from around the world to explore the acceleration of science and its applications to help solve some of humanities biggest global challenges.
Panelists tackling the issues around IP include James Donovan, OpenAI’s head of science policy and partnerships, Ken-Ichiro Nasume, Assistant Director General at the World Intellectual Property Organization ( WIPO), Antony Taubman, the former Director, Intellectual Property, Government Procurement & Competition Division of the World Trade Organization (WTO) and Christine Allan de Lavenne, an intellectual property lawyer and human rights and innovation expert.
A Growing Tension
There is a growing tension between two human rights guarantees from Article 27 of the Universal Declaration of Human Rights (UDHR): the human right to participate in and benefit from scientific progress and its applications and the right of creators, including scientists and investors, to protect their work, says GESDA’s Gerard Escher, a biologist by training. These human rights have historically co-existed but as scientific discovery accelerates through technologies such as AI, quantum computing and biotechnology, the there is an opportunity to reexamine the tensions between these rights, he says.
Emerging technologies could worsen existing power imbalances by enabling major corporations, particularly in sectors like pharmaceuticals, to innovate more quickly, secure more patents and exert greater control over scientific knowledge. This raises critical questions around access, equity and fairness. How can science benefit all, including developing nations and marginalized communities when intellectual property laws often seem to lock away innovations. The panel will explore the challenges, opportunities and potential new models for IP that ensure that scientific discovery remains incentivized (a duty of the human right to science) while ensuring global access to life-saving advancements, especially in fields like healthcare and environmental technology.
While a lot of attention is given to how to best leverage AI, the world is still not fully leveraging human advances in science, says Escher. A recent article on Science.org points out that inventors often don’t credit public research, so they don’t have to list a co-inventor. “With AI, inclusion of all would become a necessary routine,” says Escher, since AI is trained on publicly available data.
“There is an opportunity to rethink the system,” says Escher. “There is a chance to break down the silos in science and include small labs with experimental research. This could be a chance to share better, to share differently, and benefit from scientific advances in new ways.”
The issues are not limited to AI. In the field of quantum computing if a company discovers a way to advance the field “they hide it to squash the competition,” he says. “There is a return of the commercial secret before the patent, and it is really terrible for innovation.”
Going forward could a new way be developed for inventors to securely deposit what they have developed to advance the science or technology and allow others to use it for the benefit of society while still profiting from their invention?
“The human right to science part of the equation has been in the sleeping beauty stage for too long,” says Monika Plozza, GESDA’s Advisor on the Human Right to Science and Lecturer at the University of Lucerne. “In recent decades IP laws have steadily expanded, increasingly favoring protection and exclusion. IP interests have been successfully implemented, as states and private actors have invested more time and resources into developing these legal frameworks, while the human right to science remains neglected, under-developed and underused.” It’s time to change that, she says.
GESDA is actively bringing all the stakeholders around the table to find a path forward. But it is clear from interviews with the panelists that there is not yet a consensus on how to do that.
Tech Is Testing The Boundaries Of The IP System
“Every time there is a technological disruption there is pressure to rewrite the rules and rebalance the intellectual property system,” says panelists Taubman, the former WTO official. “It was the case with the introduction of photography, sound recording, the Internet and digital music, he says. “The reality is that the rule makers will never catch up and anticipate the changes. With AI we are seeing that in spades now,” he says. It is a natural cyclical upheaval of the IP system, he says, “but change never happened spontaneously. It always results in hard core litigation that tests the boundaries of the existing system.”
Technology is stretching those boundaries on multiple fronts. Up until about 20 years ago all the content that required licensing was embedded in the physical world and was easier to track and understand. What’s more the current IP system, where everything is defined and managed nationally, makes less and less sense in a digital borderless world and AI is rapidly changing the role of humans as creators and inventors. “The idea that the original artist or creator is a sophisticated computer program undermines the very tenants of the IP system,” says Taubman. “The debate we see now is a healthy response to the systemic challenges.”
The Challenges Ahead
WIPO’s Natsume says his organization has been grappling with those challenges since 2019. It convenes member states and all the emerging stakeholders, including companies, individuals and academia, twice a year to discuss how to deal with the impact of technology on IP systems and the question of how to balance the human right to benefit from scientific advances and the right for innovators and creators to protect their work. At the last meeting the focus was on GenAI and how the data training is handled. At the next meeting, in November, the discussion will center on how the patent system should deal with GenAI’s output.
By way of example, under the current IP framework so called ” obvious innovations”, where the inventive contribution made is not very big, should not benefit from patent protection. But if GenAI can produce a huge amount of inventive ideas in a few seconds “how do we evaluate if a step is big enough?” to merit a patent, asks Natsume. And if AI is trained on other people’s data, how should those people be compensated for the use of their data and how do you determine how much of an idea is derivative and who should own the rights to a resulting invention?
One idea being discussed outside of WIPO could be creating a universal basic income to compensate the entire global population for its contribution to training AI, says GESDA’s Plozza.
Many challenges to the current system have no clear answer, says Natsume. “We are at the stage of considering different views,” he says.
Finding common ground will not be easy. For example, on the question of whether AI-generated artwork can be protected by copyright, the U.S. appears to take a more restrictive view than China.
And so far, courts have refused to grant patents to inventions made by either machines or algorithms. The leading patent case is in the name of Thaler (the human) and DABUS (the claimed AI ‘inventor’). It concerns two patent applications filed in 2018 by Dr Stephen Thaler at the UK’s intellectual property office, where he asserted that the inventions were created by an AI machine called DABUS without the involvement of a human inventor. Patent experts like Taubman think governments will eventually need to weigh in and that some form of protection will ultimately be granted to non-human inventors.
Is IP Dead?
Due to accelerated scientific progress and new tech developments, there is a need for changes in the current IP framework, says IP lawyer Allan de Lavenne, “but we won’t ditch IP altogether, IP is not dead.” IP and the right to science “are not the enemies they are often claimed to be, but rather have beneficial interactions/areas,” she says. There are three main “dials” in current IP law: exclusions, exceptions and expiration that, through better cooperation between human rights and intellectual property actors, can be adjusted to find the new balance between IP and the right to science we need, she says.
She points out that there are already exceptions to patent law that place certain advances outside the scope of protection for the purpose of scientific progress as well as exceptions allowing for them to benefit the largest number of people. Compulsory licenses for vaccines were granted during COVID and will continue, with adjustments, to be issued, she says. The 20 years duration of patent protection might need to be rethought, as well as practices such as improvement patents which end up granting protection forever, she says, “but overall IP law has the tools we need to make the adjustments.”
Defining The Way Forward
Taubman, the former WTO official, says it is important to remember that the solutions don’t have to be binary. “The future should not be limited to an extension of the current patent system or nothing,” he says. Perhaps a “patent-like” tool could be created that would still include economic incentives and recognition to this new category of AI ‘inventors’ without granting exclusive rights. “Changing innovation models may lead us to rethink the way drug discovery is financed,” he says. “But if we don’t want exclusive rights who is going to pay? Early innovation is one thing, but drug development will remain costly and risky. Governments and philanthropic initiatives can complement private sector players. New models need to be discussed and explored.” In a world of accelerating AI and scientific discoveries, says Taubman, “we can’t afford to be stuck with one narrow innovation model.”
This article is content that would normally only be available to subscribers. Sign up for a four-week free trial to see what you have been missing.
To access more of The Innovator’s Focus On AI articles click here.
