Patent Law For Machine-Intelligent Biotechnology.

01 Apr 2026 --
0 Comments

1. Introduction: Patent Law and Machine-Intelligent Biotechnology

Machine-Intelligent Biotechnology (MIB) refers to innovations at the intersection of AI/ML technologies and biotech, such as:

AI-designed proteins, enzymes, or genetic sequences
Machine-learning-based drug discovery
AI-assisted diagnostic tools using biomarkers
Bioinformatics algorithms for genomic predictions

Patentability Criteria (Generally under TRIPS, 35 U.S.C., or Indian Patents Act 1970):

Novelty (New Invention) – Must not be disclosed anywhere in the world before filing.
Inventive Step / Non-Obviousness – Must not be obvious to a person skilled in the art.
Industrial Applicability / Utility – Must be capable of practical application.
Subject Matter Eligibility – Certain natural phenomena, abstract ideas, or algorithms alone are not patentable.

The tension arises because MIB inventions often involve AI algorithms (potentially abstract ideas) and biological materials (sometimes considered naturally occurring).

2. Key Issues in Patenting Machine-Intelligent Biotechnology

AI-generated inventions: Who is the inventor? AI itself? Human programmer?
Patentable subject matter: Can an algorithm-derived biological sequence or protein be patented?
Disclosure & Enablement: How much detail is required if AI generates unpredictable results?
Obviousness / Inventive Step: If AI predicts a protein structure, is it obvious to someone in biotech?
Ethical & policy constraints: Many jurisdictions prohibit patenting certain life forms or genes in natural form.

3. Landmark Cases

Case 1: Diamond v. Chakrabarty, 447 U.S. 303 (1980) – U.S.

Facts: A genetically modified bacterium capable of breaking down crude oil was created by Dr. Chakrabarty. The U.S. Patent Office initially rejected it, claiming living organisms are not patentable.
Decision: The Supreme Court held that “anything under the sun made by man” is patentable, including genetically modified organisms.
Significance for MIB:
- Opens the door for patenting AI-designed microorganisms or synthetic life forms.
- Shows that human ingenuity (even when assisted by AI) can qualify as “man-made.”

Case 2: Myriad Genetics, Inc. v. Association for Molecular Pathology, 569 U.S. 576 (2013) – U.S.

Facts: Myriad discovered the BRCA1 and BRCA2 genes associated with breast cancer and sought patents on the isolated genes.
Decision: The Supreme Court ruled naturally occurring DNA is not patentable, but cDNA (synthetic) is patentable.
Significance for MIB:
- AI can design synthetic genes or proteins, which can be patentable.
- Reinforces the boundary: natural sequences discovered by AI may not be patentable, but engineered ones are.

Case 3: Thaler v. USPTO – DABUS AI Inventor (2021) – U.S.

Facts: Stephen Thaler claimed that his AI system “DABUS” independently invented new inventions. The patent application listed DABUS as the inventor.
Decision: The USPTO rejected the claim; only natural persons can be inventors.
Significance for MIB:
- AI cannot currently be listed as an inventor.
- In biotech AI inventions, the human programmer or operator must be credited.
- Raises questions on IP ownership for AI-driven discoveries in drug design or genomics.

Case 4: Eli Lilly v. Canada (2017) – Canada

Facts: Eli Lilly challenged Canada’s rejection of certain pharmaceutical patents.
Decision: Canadian courts emphasized utility and promise doctrine – inventions must deliver what is promised in the patent.
Significance for MIB:
- AI-derived biotech inventions must demonstrate practical applicability.
- Hypothetical predictions by AI are not enough; experimental validation strengthens patent claims.

Case 5: Regeneron Pharmaceuticals v. Merus (2019) – U.S.

Facts: Dispute over antibody sequences generated through computational methods.
Decision: Court held that AI-designed antibodies could be patentable if human involvement and non-obvious modifications exist.
Significance for MIB:
- Confirms patentability of AI-assisted biotech innovations.
- Human ingenuity combined with AI predictions satisfies the inventive step requirement.

Case 6: DeepMind/AlphaFold Implications (Hypothetical Legal Debate)

Context: DeepMind’s AlphaFold predicts protein structures using AI.
Legal Questions:
- Are AI-predicted protein structures inventions or discoveries?
- Does filing a patent require lab synthesis or validation?
Implications:
- AI-assisted predictions alone may not be enough; a human-guided experiment or synthesis increases patent eligibility.
- Jurisdictions may vary: the EU is stricter on “abstract computational results” vs. U.S. leaning toward “practical applications.”

4. Practical Guidelines for MIB Patenting

Human Inventor Attribution: Always list a natural person, not AI.
Synthetic vs. Natural Sequences: Patent synthetic genes, proteins, or AI-modified organisms, not naturally occurring sequences.
Demonstrate Utility: Validate AI-predicted outcomes in the lab.
Include Algorithm Disclosure Carefully: Focus on how AI leads to biotech invention, not the AI algorithm itself.
Document Human Contribution: Highlight human creativity, intervention, or guidance in generating the invention.

5. Summary Table of Cases

Case	Jurisdiction	Key Takeaway for MIB
Diamond v. Chakrabarty (1980)	U.S.	Human-made life forms patentable
Myriad Genetics (2013)	U.S.	Natural genes not patentable, synthetic ones are
Thaler/DABUS (2021)	U.S.	AI cannot be listed as inventor
Eli Lilly v. Canada (2017)	Canada	Must demonstrate practical utility
Regeneron v. Merus (2019)	U.S.	AI-designed biotech inventions patentable with human involvement
AlphaFold debate	Global	AI predictions alone insufficient; human verification recommended