By Hélio Roque, European Patent Attorney, Biotech & Life Sciences Department, ABG IP
On November 30, 2022, the world was introduced to ChatGPT—a breakthrough that redefined how we interact with machines, devices, and, most importantly, data. The floodgates opened, and adoption was swift. Headlines quickly proclaimed that artificial intelligence would transform our lives, reshape jobs, and leave a lasting impact on society. For simplicity, let’s call this new era, beginning in late 2022, the “AI boom.”
One of the most discussed consequences of the AI boom, almost from day one, has been its transformative effect on the patent ecosystem. This ecosystem spans two dimensions: the entire patent application process—from filing to expiration, including office actions, oppositions, litigation, and licensing—and the way patent professionals, such as attorneys and consultants, work. News on these changes has circulated widely, from debates over whether AI can be considered an inventor[1] to predictions of a surge in patent filings driven by generative AI’s ability to accelerate innovation across technical fields, as highlighted in a recent Duke Law report (Rai, A., Duke Law, 2025)[2].
While shifts in professional workflows and procedural norms are harder to quantify in the short term, tangible evidence of the AI boom’s impact is emerging in biotechnology. A clear example is the U.S. Food and Drug Administration’s recently launched webpage listing approved AI-enabled medical devices. The data shows an exponential rise in such devices since 2015, with a sharp spike in 2023 and continued growth through 2024 and 2025.
Figure 1: Number by years of FDA AI-Enabled Medical Devices authorized for marketing in the United States[3].
Three years into the “AI boom,” a key question arises: Are we seeing a surge in patent applications at the European Patent Office (EPO), particularly in biotechnology and pharmaceuticals—the sectors most relevant to my work?
To explore this, I turned to the EPO’s Statistical Center to examine whether there has been a noticeable spike in new patent filings. My analysis covered both overall application trends and specific fields: Computer Science, Biotechnology, and Pharmaceuticals. While my primary interest lies in the latter two, the pervasive nature of AI across all technical domains made it essential to also review Computer Science, where most AI-related inventions are likely concentrated.
I focused on European patent applications originating from the “Big Five” patent offices—the EPO (Europe), USPTO (United States), JPO (Japan), KIPO (South Korea), and CNIPA (China)—tracking trends from 2016 onward. This timeframe accounts for the likelihood that companies and developers were already leveraging AI before its public debut in late 2022 and allows to visualize if any trend predates the AI boom.
Figure 2: Overall European patent applications from applicants from the Big Five.
Figure 2 reveals a steady rise in European patent applications from EPO-based applicants over recent years. In contrast, Chinese applicants show a sharper growth rate, while filings from the United States, South Korea, and Japan have remained relatively stable.
Figure 2 indicates no significant increase in overall European patent applications between 2020 and 2025. This suggests that, if an AI-driven surge exists, it has yet to impact aggregate numbers—or is confined to specific technological fields without influencing the broader trend. To investigate further, I examined application volumes by technology sector, starting with computer technology, to see whether any noticeable uptick could be detected.
Figure 3: European patent applications in the field of computer technology from applicants of the Big Five.
Figure 3 illustrates notable trends in European patent applications related to computer technology. From 2020 onward, filings from European applicants accelerated at a faster pace compared to previous years. For U.S. applicants, growth appears to have occurred earlier—between 2018 and 2019—maintaining momentum until 2022–2023, when a dip was observed, followed by a return to the prior growth rate.
Chinese applicants recorded the most significant surge between 2020 and 2021, though this rapid growth plateaued in 2023 and slightly declined in 2024, which returned to growth in 2025. South Korean filings also spiked between 2022 and 2023, while applications from Japan remained largely stable throughout the analyzed period.
Overall, Figure 3 suggests heightened activity in computer technology since 2019, as evidenced by the upward trend in applications. This pattern may reflect the influence of the AI boom, correlating with the surge of innovation and foundational shifts driven by AI in computing and related fields.
And what about Biotechnology and Pharma?
From Figure 1, it is evident that AI is already making a significant impact on medical devices and shaping clinical procedures. Scientific literature reinforces this trend, highlighting extensive work in areas such as target discovery, compound screening, medical data analysis, and diagnostics. In particular, AI is transforming drug development by unlocking new insights from complex multi-omics biomedical data. But does this activity translate into a measurable rise in patent filings?
Figure 4 shows that European patent applications from EPO-based applicants had a dip from 2020 to 2021, followed by an increased, though the growth has been modest, with the number of applications more or less equal for the last 3 years. Applications from U.S. and Chinese applicants, by contrast, have grown steadily over the same period, though in the last year the number of applications from the US decreased. Meanwhile, filings from Japan and South Korea have risen at a modest pace.
Figure 4: European patent applications in the field of Biotechnology from applicants of the Big Five.
For both European and U.S. applicants, the upward shift in patent filings appears to have begun around 2016–2018—well before the AI boom—likely reflecting the rise of other groundbreaking technologies in biotechnology, such as CRISPR-Cas9 and novel cancer immunotherapies.
In the pharmaceutical sector, Figure 5 reveals a distinct trend compared to other fields. After a period of steady growth in European patent applications from 2016 through 2021–2022 for U.S., Chinese, and European applicants, filings plateaued around 2022 and then declined from 2023 to 2025, with a sharper decline in Us patent application. This suggests that, at least in Pharma, the AI boom has not translated into an increase in European patent applications, and if anything appears to be correlated with a decrease in European patent applications.
Figure 5: European patent applications in the field of Pharma from applicants of the Big Five.
Overall, two years after the onset of the AI boom, no major effect on European patent filings is evident—neither in aggregate nor in the specific fields of biotechnology and pharmaceuticals. The only notable growth appears in computer technology, and even there, the increase has been modest.
This raises an important question: Why aren’t we seeing a surge in patent applications from a technology hailed as revolutionary? The simplest answer might be a disconnect between AI’s promise and its current capabilities. In that case, the only option is to wait and see what AI delivers in the coming years.
A more realistic explanation, however, is that AI has yet to fully mature. Many real-world applications are still under development and have not reached the stage where they can generate viable patent filings—particularly in biotechnology and pharma. While some solutions, such as AI-powered diagnostic tools that detect previously invisible patterns in patient data (gene expression, imaging, physical metrics), may lead to immediate patentable outcomes, other aspects of the AI boom do not produce such direct, patent-ready results.
Take drug development as an example. Figure 6 outlines the typical pipeline in the pharmaceutical industry: from target identification to lead optimization, followed by preclinical testing and clinical trials. Each stage is complex and time-consuming, meaning that even if AI accelerates early discovery, the impact on patent filings may take years to materialize.
Figure 6: General pipeline diagram of pharma drug discovery (AI generated).
As noted earlier, AI is transforming the drug discovery pipeline, particularly in the early stages. By leveraging advanced analytics across multiple omics datasets and pathway analyses, AI is uncovering new targets and identifying potential compounds that were previously undetectable. It also enables in-silico testing of compound effects on existing and novel targets, rapidly generating promising candidates for therapeutic development.
However, beyond these initial steps, AI’s ability to accelerate progress is almost inexistent. Preclinical testing—the stage at which most patent applications for new drugs are filed—remains a critical bottleneck. This phase is essential for substantiating medical claims and a patent application, and justifying the significant investment required for a clinical trial. While AI has sped up target identification and lead optimization, the pace of preclinical and clinical testing has not changed, exacerbating an already existing constraint.
Drug development has long been limited by these later stages, even after advances such as automated screening and the availability of omics data. In vivo testing remains slow—perhaps necessarily so—because, as is well known, in vitro results often fail to translate into desired effects in living systems.
The challenge for biotechnology and pharma may lie in the widening gap between rapid, AI-driven data analysis and the slower generation of hard experimental evidence. While AI makes it easier to propose new hypotheses and insights, the capacity to validate them has not kept pace. This imbalance may explain why, at least for now, we have not seen a surge of European patent applications in these fields despite the AI boom.
In sum, while AI’s promise remains real, its impact on patent filings will likely unfold gradually, as the technology matures and integrates deeper into processes that require rigorous validation. For now, the revolution is happening in theory and early discovery—not yet in the patent receiving office.
