Ipr In AI-Assisted Medical Diagnostics.
IPR in AI-Assisted Medical Diagnostics
1. Introduction
AI-assisted medical diagnostics use algorithms (often machine learning or deep learning) to detect, predict, or classify diseases from medical data such as X-rays, MRIs, CT scans, pathology slides, ECGs, and patient records. Examples include AI systems for cancer detection, diabetic retinopathy screening, and cardiac risk prediction.
These systems raise complex Intellectual Property Rights (IPR) issues because:
AI systems rely on software algorithms
They are trained on medical data
They often improve autonomously
Their outputs may affect human life, invoking regulatory and ethical concerns
The major IPR regimes involved are:
Patent law
Copyright law
Trade secrets
Data rights
Ownership and inventorship rules
2. Patent Law Issues in AI Medical Diagnostics
Key Challenges
Whether AI-based diagnostic methods are patentable subject matter
Whether algorithms constitute abstract ideas
Who qualifies as the inventor — human or AI?
Patentability of medical methods
3. Landmark Case Laws (Detailed Explanation)
Case 1: Association for Molecular Pathology v. Myriad Genetics, Inc. (2013, USA)
Facts
Myriad Genetics discovered the location of BRCA1 and BRCA2 genes, mutations of which increase breast and ovarian cancer risk. Myriad patented:
Isolated DNA sequences
Diagnostic methods for detecting cancer risk
Issue
Whether naturally occurring genetic material and diagnostic methods can be patented.
Judgment
The US Supreme Court held:
Naturally occurring DNA is not patentable
Synthetic DNA (cDNA) may be patentable
Diagnostic methods based purely on natural phenomena are not patent-eligible
Relevance to AI Diagnostics
AI diagnostic tools often detect natural biological correlations
Merely identifying patterns in medical data may be considered a law of nature
AI diagnostic patents must include technical innovation, not just discovery
Impact
AI medical diagnostic inventions must:
Go beyond data interpretation
Include novel technical processes
Show human intervention and technological improvement
Case 2: Mayo Collaborative Services v. Prometheus Laboratories (2012, USA)
Facts
Prometheus patented a method to:
Measure metabolite levels in blood
Adjust drug dosage accordingly
Issue
Whether diagnostic methods based on natural correlations are patentable.
Judgment
The Supreme Court ruled:
The method merely applied a natural law
Steps involved were routine and conventional
Patent was invalid
AI Relevance
AI diagnostic tools often:
Correlate biomarkers with disease outcomes
Provide treatment recommendations
If AI only automates a doctor’s mental process, it may be:
Considered an abstract idea
Unpatentable
Principle Established
To be patentable, AI diagnostics must add:
A novel technical step
A specific technological improvement
Case 3: Alice Corp. v. CLS Bank International (2014, USA)
Facts
Alice Corp patented a computer-implemented method for financial transactions.
Issue
Whether computer-implemented inventions are patentable.
Judgment
The Court introduced the Alice Test:
Is the claim directed to an abstract idea?
Does it include an “inventive concept” that transforms it into patent-eligible subject matter?
Application to AI Diagnostics
Most AI medical diagnostic patents face rejection under Alice because:
Algorithms are considered abstract ideas
Merely running them on a computer is insufficient
Implication
AI diagnostic inventions must:
Improve computer functionality
Improve medical imaging processing
Show technical architecture innovation
Case 4: Thaler v. Commissioner of Patents (DABUS Case, 2021–2023)
Facts
Stephen Thaler filed patent applications listing DABUS, an AI system, as the inventor.
Issue
Can an AI system be named as an inventor?
Judgment
Courts in multiple jurisdictions held:
Inventor must be a natural person
AI cannot hold legal rights or responsibilities
Impact on AI Medical Diagnostics
AI-generated diagnostic innovations cannot claim AI as inventor
Ownership lies with:
Developers
Employers
Research institutions
Significance
Reinforces the human-centric nature of patent law, even in advanced AI systems used in medicine.
Case 5: Vanda Pharmaceuticals Inc. v. West-Ward Pharmaceuticals (2018, USA)
Facts
Vanda patented a method using genetic testing to determine drug dosage for schizophrenia patients.
Issue
Whether personalized medical treatment methods are patentable.
Judgment
The Court upheld the patent because:
It involved a specific treatment step
It went beyond natural correlations
It applied results in a concrete medical action
Importance for AI Diagnostics
AI systems that:
Diagnose disease and
Direct specific treatment actions
are more likely to be patent-eligible.
Case 6: Indian Scenario – Computer Related Inventions (CRI) and AI Diagnostics
Indian Patent Office Position
India excludes:
Algorithms per se
Mathematical methods
Methods of treatment
However, patents may be granted if:
AI is tied to hardware
Produces technical effect
Improves diagnostic equipment performance
Example Applications
AI-enabled imaging machines
AI-based ECG interpretation systems
AI-assisted pathology scanners
Practical Outcome
Many AI medical diagnostic tools in India are protected through:
Trade secrets
Copyright in software
Data exclusivity
4. Copyright Issues
Key Question
Can AI-generated diagnostic reports be copyrighted?
Legal Position
Copyright requires human authorship
AI-generated outputs usually lack originality
Copyright may subsist in:
Software code
Training datasets (if curated creatively)
User-modified reports
Ownership
Typically vests in:
Software developers
Hospitals (work-for-hire)
AI service providers
5. Trade Secrets and Data Protection
Importance
Training datasets in AI diagnostics are extremely valuable.
Protection Includes:
Proprietary datasets
Model architecture
Training methods
Hospitals often rely on:
Confidentiality agreements
Licensing contracts
6. Liability and Ownership Issues
Who is Responsible for Wrong Diagnosis?
AI developer?
Hospital?
Doctor using AI?
Courts currently lean toward:
Human oversight responsibility
AI as a decision-support tool
7. Conclusion
AI-assisted medical diagnostics challenge traditional IPR frameworks. Courts worldwide have adopted a cautious approach, emphasizing:
Human inventorship
Technical innovation
Protection of natural phenomena
RELATED Blog
comments