Introduction
As biotechnology and artificial intelligence (AI) evolve at an unprecedented pace, they continue to challenge traditional legal concepts of patentability. Central to this tension is a key question: where do we draw the line between invention and discovery?
This line determines what can—and cannot—be protected by a patent. Courts across different jurisdictions are being asked to decide whether a gene sequence, a diagnostic method, or an AI-generated invention qualifies as a human-made invention, or merely a natural discovery or abstract idea.
In this article, we explore recent court decisions shaping this boundary, analyze key legal principles, and highlight the practical implications for innovators in biotech and AI.
1. The Legal Framework: What Is Patentable?
In general, for an invention to be patentable, it must meet four main criteria:
- Novelty
- Inventive step (non-obviousness)
- Industrial applicability (usefulness)
- Patentable subject matter
The last criterion—what constitutes patentable subject matter—is where biotech and AI often run into legal uncertainty.
Biotech Example: Natural DNA Sequences
A classic case is Association for Molecular Pathology v. Myriad Genetics (U.S., 2013). Myriad had isolated the BRCA1 and BRCA2 genes, mutations of which are linked to breast cancer, and patented the isolated sequences.
The U.S. Supreme Court ruled that naturally occurring DNA sequences are not patentable because they are products of nature. However, complementary DNA (cDNA)—synthetically created and not naturally occurring—was deemed patentable.
🔍 Takeaway: Discovery of something that exists in nature, even if useful, may not qualify as an invention.
2. Artificial Intelligence and the “Inventor” Question
The rise of AI has introduced a more abstract challenge: Can an AI system be an inventor?
Case: DABUS Patent Applications (UK, US, EU, Australia, South Africa)
Dr. Stephen Thaler filed patents naming DABUS, an AI system he developed, as the inventor of two innovations: a food container and a warning light. Patent offices in the UK, US, and EU rejected the application, reasoning that:
- Only natural persons can be inventors under current law.
- AI lacks legal personhood and therefore cannot hold inventorship.
However, in a surprising turn, South Africa granted the patent, recognizing DABUS as the inventor—the first country to do so. Similarly, an Australian court initially ruled in favor of DABUS, though the decision was later overturned on appeal.
🧠 Key Insight: Legal systems are grappling with whether human involvement is necessary for an invention to be patentable.
3. Diagnostic Methods and the “Abstract Ideas” Problem
Diagnostic tools powered by AI or biotech also face patent hurdles. Many courts reject such patents on the grounds that they represent abstract ideas or laws of nature, particularly in the U.S.
Example: Mayo Collaborative Services v. Prometheus Laboratories (U.S., 2012)
The Supreme Court invalidated a patent for a diagnostic method that correlated drug metabolite levels with the effectiveness of a treatment. The reasoning? The method was essentially a natural law plus routine steps.
⚖️ Lesson: Even when AI or biotech tools provide valuable insights, if the underlying principle is a natural correlation or abstract idea, it may not be patentable.
4. The European Approach: Technical Character Matters
In Europe, patentability is less about the origin (e.g., nature vs. invention) and more about the technical character of the claim.
For instance:
- The European Patent Office (EPO) allows biotech patents if a biological material has been isolated and is industrially applicable.
- For AI, algorithms per se are not patentable, but if an AI contributes to a technical solution, it can be.
Example: EPO Guidelines on AI
A machine learning algorithm used to improve image processing in medical imaging could be patentable because it solves a technical problem.
🧩 Insight: The EPO’s more pragmatic, case-by-case approach provides more flexibility—but still requires a clear technical contribution.
5. Global Trends and Strategic Considerations
Given the divergence in how courts treat biotech and AI inventions, innovators and patent attorneys must:
- Tailor applications to national laws
- Emphasize human involvement or technical effect
- Avoid claims that rely solely on natural laws or abstract concepts
Multinational patent strategies may include claim drafting tailored for specific jurisdictions, or filing in more AI-friendly countries like South Korea, China, or South Africa.
Conclusion
The distinction between invention and discovery is not just philosophical—it’s legal and practical. As biotech decodes nature and AI begins to mimic human creativity, the challenge for courts is to ensure the patent system both rewards innovation and respects legal boundaries.
Innovators must be aware of these boundaries and adapt their patent strategies accordingly. Understanding how courts are interpreting the law today will shape how we protect the inventions of tomorrow.
Further Reading & References
- Association for Molecular Pathology v. Myriad Genetics, 569 U.S. 576 (2013)
- Mayo Collaborative Services v. Prometheus Laboratories, 566 U.S. 66 (2012)
- Thaler v. Commissioner of Patents [2021] FCA 879 (Australia)
- EPO Guidelines for Examination – G-II, 3.3 (Artificial Intelligence)
- World Intellectual Property Organization (WIPO) – “WIPO Technology Trends 2019: Artificial Intelligence”
- European Parliament Briefing – “Artificial Intelligence and Civil Law” (2020)