The Trump administration has quietly reversed a key Biden-era policy on artificial intelligence and patent law, signaling that AI systems can now be treated as mere tools — like a calculator — rather than as potential co-inventors. The shift, which rescinds previous guidance on when a human might qualify as a co-inventor of an AI-assisted discovery, has profound implications for the pharmaceutical industry, where AI is increasingly used to design novel molecules and accelerate drug development. At stake is nothing less than the future of innovation in an industry where the difference between a breakthrough therapy and a dead end can hinge on the legal status of a machine’s contribution.
The Constitutional Question: What Does It Mean to Invent?
The core of the debate rests on the very purpose of intellectual property law. As policy expert Abbott notes, the point of our intellectual-property laws is to encourage innovation. This principle is enshrined in Article 1 of the US Constitution, which grants exclusive rights to inventors and authors for a limited time precisely “to promote the Progress of Science and the Useful Arts.” The question now confronting the patent office is whether an AI system can be said to “promote progress” in the same way a human inventor does, and whether the exclusion of AI-generated outputs from patent protection would undermine that constitutional mandate.
Under the Biden administration, the US Patent and Trademark Office (USPTO) published detailed guidance to help applicants determine whether and when a human would truly qualify as a co-inventor of an AI-assisted discovery. The guidance was intended to provide clarity in a rapidly evolving field, where AI systems are not merely crunching numbers but are generating novel chemical structures and biological insights that no human had previously conceived. The Biden-era approach acknowledged that AI could play a significant role in the inventive process, while still requiring a meaningful human contribution for a patent to be valid.
That approach has now been rescinded. The Trump administration’s patent office now says that AI is merely a tool, no different from a calculator or a microscope, and that there is no need to even mention its use in a patent application. This don’t-ask-don’t-tell approach, as some observers have dubbed it, effectively sidesteps the difficult questions about AI inventorship by pretending they do not exist.
The Patent Office’s Reversal: From Guidance to Silence
The reversal is not a formal rule change but a shift in policy guidance. The Biden-era guidance was intended to help patent applicants navigate the murky waters of AI-assisted invention. It outlined factors that would indicate a human had made a “significant contribution” to the invention, even if an AI system had also played a role. The guidance was seen as a pragmatic attempt to keep the patent system relevant in an era of increasingly autonomous AI tools.
By rescinding that guidance, the Trump administration has effectively told patent applicants that they do not need to disclose AI involvement at all. This reduces the administrative burden on applicants, but it also creates uncertainty. If a patent is later challenged on the grounds that the true inventor was an AI system, the lack of disclosure could become a legal vulnerability. The patent office’s stance also raises the question of whether an invention that was entirely generated by an AI, with no meaningful human input, could be patented at all.
Currently, the US patent office seems to be taking a don’t-ask-don’t-tell approach to the use of AI. This is a stark contrast to the approach of the US Copyright Office, which is already refusing to grant copyrights to images and text generated by AI. The copyright office’s stance has raised concerns from organizations like the Motion Picture Association of America, whose members are increasingly using AI tools in film production and visual effects. The parallel between copyright and patent law is instructive: if the patent office follows the same trajectory as the copyright office, it could ultimately refuse to grant patents for AI-generated inventions, regardless of whether a human was involved in the process.
What Is the Current US Patent Office Stance on AI Inventorship?
The current US patent office stance is that AI is a tool, like a calculator, and that there is no requirement to disclose its use in a patent application. The Biden-era guidance that helped applicants determine whether a human qualified as a co-inventor of an AI-assisted discovery has been rescinded. This means that patent applicants are now free to ignore the role of AI in the inventive process, leaving the question of AI inventorship largely unaddressed by the patent office. The stance effectively creates a legal gray area where the boundaries of human and machine contribution are undefined.
The Industry Response: Keeping Humans in the Loop
Pioneering AI drug companies are not waiting for the legal system to catch up. They are keeping humans in the loop, at least for now, and documenting everything carefully. At Insilico Medicine, a company that uses AI for drug discovery, CEO Alex Zhavoronkov describes a process where human chemists still have to synthesize the drugs, create variants, and test them on animals. “That’s the person who is going to be named on the patent,” he says. “And even if you decided to completely roboticize this process, including the experiments, someone will still push the button and give the budget.”
This pragmatic approach reflects the current reality of AI drug development. AI systems can generate thousands of potential drug candidates in silico, but the process of synthesizing, testing, and optimizing those candidates still requires human expertise. The human chemist who designs the synthesis route, interprets the experimental data, and decides which variants to pursue is performing a genuinely inventive role. The question is whether that role is sufficient to qualify as a “significant contribution” under patent law, especially if the AI system was the one that originally proposed the molecular structure.
Insilico’s approach is likely to be replicated across the industry. Companies are keeping detailed records of human involvement in the drug development process, including the design of experiments, the interpretation of results, and the decision-making that leads to a patentable invention. This documentation could be crucial if a patent is later challenged on the grounds that the AI system was the true inventor.
The Thought Experiment That Keeps Legal Scholars Up at Night
Abbott poses a provocative thought experiment that cuts to the heart of the issue. “What if I asked Claude to cure cancer, and it did?” he asks. “I think it would be inappropriate to claim that I invented that.” The question is not merely hypothetical. AI systems are becoming increasingly capable of generating novel solutions to complex problems, and it is only a matter of time before an AI system proposes a drug candidate that no human would have thought of. If the only human contribution was to ask the question, does that make the person an inventor? Abbott suggests that the answer is no, and that the patent system would need to grapple with the concept of an AI inventor.
This thought experiment highlights the fundamental tension in the patent office’s current stance. By treating AI as a mere tool, the patent office is avoiding the question of whether an AI could be an inventor. But the question will not go away. As AI systems become more autonomous, the line between tool and inventor will become increasingly blurred. The patent office’s don’t-ask-don’t-tell approach may work for now, but it will eventually need to confront the issue head-on.
The Broader Context: AI and the Future of Drug Development
Abbott’s worry is that if US policy excludes AI-generated outputs from protection, that could put a damper on future drug development. The pharmaceutical industry is already facing a productivity crisis, with the cost of developing a new drug estimated at over $2 billion and the success rate of clinical trials declining. AI has the potential to dramatically reduce the cost and time of drug discovery, by screening millions of compounds in silico and identifying the most promising candidates for further development. But if the outputs of those AI systems cannot be patented, the economic incentive for pharmaceutical companies to invest in AI-driven drug discovery would be significantly diminished.
The impact could be particularly severe for smaller biotech companies, which often rely on patent protection to attract investment. A startup that uses AI to discover a novel drug candidate might struggle to secure funding if investors are uncertain whether the resulting patents will be enforceable. Larger pharmaceutical companies, with their deep pockets and extensive patent portfolios, might be better positioned to absorb the risk, but even they would be reluctant to invest billions of dollars in a drug that could be copied by competitors as soon as it is approved.
The US Copyright Office’s refusal to grant copyrights to AI-generated images and text provides a cautionary precedent. The Motion Picture Association of America has expressed concerns that the copyright office’s stance could stifle innovation in the film industry, where AI tools are increasingly used for visual effects, animation, and even scriptwriting. If the patent office follows a similar path, the pharmaceutical industry could face a similar disruption.
The Legal Landscape: What Comes Next
The question of AI inventorship is likely to be resolved in the courts, not in the patent office. Abbott says that the question of whether “pushing a button” counts as being an inventor is a question for future legal cases. The patent office’s current stance provides little guidance, leaving the courts to interpret the law in the context of a rapidly evolving technology.
Several key legal questions are likely to arise. First, what constitutes a “significant contribution” to an invention when an AI system is involved? The Biden-era guidance attempted to answer this question, but the rescission of that guidance has left a vacuum. Second, what is the legal status of an invention that was entirely generated by an AI, with no meaningful human input? The current patent law requires that an inventor be a “natural person,” but the definition of a natural person may need to be reconsidered in the age of AI. Third, what is the liability of a patent applicant who fails to disclose the use of AI in the inventive process? The don’t-ask-don’t-tell approach may create legal risks for applicants who are not transparent about the role of AI.
The international dimension adds another layer of complexity. The European Patent Office and the UK Intellectual Property Office have both taken the position that an AI system cannot be an inventor, but they have also provided guidance on how to handle AI-assisted inventions. The US patent office’s reversal could put the US out of step with its international partners, creating confusion for multinational pharmaceutical companies that seek patent protection in multiple jurisdictions.
The Practical Implications for Drug Companies
For pharmaceutical companies that are already using AI in drug discovery, the practical implications of the patent office’s reversal are mixed. On the one hand, the removal of the guidance reduces the administrative burden of patent applications. Companies no longer need to document the specific role of AI in the inventive process, which could simplify the patent application process. On the other hand, the lack of guidance creates uncertainty about the validity of patents that involve AI. A company that obtains a patent under the current don’t-ask-don’t-tell regime could face a challenge later if a competitor argues that the true inventor was an AI system.
Companies like Insilico are taking a cautious approach, keeping humans in the loop and documenting everything carefully. This is likely to be the prudent strategy for the foreseeable future. By ensuring that human chemists are involved in the synthesis, testing, and optimization of drug candidates, and by documenting that involvement, companies can strengthen their claims to inventorship. The human who synthesizes the drug, creates variants, and tests them on animals is clearly making a significant contribution, even if the AI system was the one that originally proposed the molecular structure.
But the question of what happens when the entire process is automated remains. As Zhavoronkov notes, even if the process is completely roboticized, someone will still push the button and give the budget. The question is whether pushing that button counts as an inventive contribution. Abbott suggests that it does not, and that the legal system will need to develop a new framework for assessing AI inventorship.
The Strategic Significance of the Reversal
The Trump administration’s reversal of the Biden-era AI patent guidance is more than a technical policy change. It is a strategic signal about the role of AI in the economy and the relationship between human and machine intelligence. By treating AI as a mere tool, the administration is taking a conservative approach that prioritizes human agency and control. This approach is consistent with the broader deregulatory agenda of the Trump administration, which has sought to reduce the regulatory burden on businesses and encourage innovation.
But the conservative approach may also stifle innovation in the long run. By refusing to acknowledge the role of AI in the inventive process, the patent office is creating a legal environment that is uncertain and potentially hostile to AI-driven innovation. Companies that are developing AI tools for drug discovery may be reluctant to invest in the technology if they are uncertain about the patentability of the outputs. The result could be a slowdown in the development of AI-driven therapies, which would be a significant loss for patients and for society.
The timing of the reversal is also significant. The pharmaceutical industry is in the midst of a transformation, with AI tools becoming increasingly sophisticated and capable. The patent office’s stance could shape the trajectory of that transformation, by determining whether the economic rewards of AI-driven drug discovery accrue to the companies that develop the AI tools, the companies that use the tools, or the public domain. The outcome of this policy debate will have lasting implications for the pharmaceutical industry and for the broader innovation economy.
The patent office’s reversal also raises questions about the role of the US in the global AI landscape. The US has long been a leader in both AI technology and pharmaceutical innovation. The patent office’s stance on AI inventorship could affect the competitiveness of the US pharmaceutical industry, as companies may choose to develop and patent AI-driven drugs in jurisdictions with more favorable legal frameworks. The European Union, for example, is developing a comprehensive regulatory framework for AI that could provide more clarity on the issue of AI inventorship.
Ultimately, the question of AI inventorship is not just a legal question; it is a philosophical question about the nature of invention and the role of human creativity. The US Constitution’s mandate to “promote the Progress of Science and the Useful Arts” was written in an era when the only conceivable inventors were human beings. Today, we are confronting the possibility that machines might be capable of inventing things that no human would have thought of. The patent system, which was designed to encourage human innovation, must now grapple with the question of whether and how to encourage machine innovation. The Trump administration’s reversal is a step in that debate, but it is far from the final word. The courts, the Congress, and the international community will all have a role to play in shaping the future of AI inventorship. And the pharmaceutical companies that are investing in AI-driven drug discovery will be watching closely, because the outcome of this debate will determine the value of their investments and the future of their industry.
