Ipr In AI-Assisted Clinical Decision Support Systems Ip.
I. Intellectual Property in AI-Assisted Clinical Decision Support Systems
AI-Assisted Clinical Decision Support Systems (AI-CDSS) are software tools that help clinicians analyze patient data, interpret lab results, suggest diagnoses, or recommend treatments. They often rely on machine learning algorithms, large datasets, and integrations with electronic health records (EHRs).
Key IP issues in AI-CDSS include:
Patent eligibility – Can an AI-based clinical method or software be patented?
Inventorship and ownership – Who owns inventions created with AI?
Scope of protection – What aspects of AI-CDSS can be protected?
Patent enforcement – How are disputes resolved in courts?
II. Important Patent Law Principles for AI-CDSS
Patentable Subject Matter
Laws of nature, natural phenomena, and abstract ideas are not patentable.
Courts use a two-step test:
Is the claim directed to a law of nature, abstract idea, or natural phenomenon?
If yes, does it include an inventive concept that transforms it into patentable subject matter?
Technical Innovation Required
Simply applying AI to medical data is often insufficient.
Courts prefer inventions that improve computing methods, speed, accuracy, or workflow integration.
Human Inventorship
AI cannot legally be recognized as an inventor. A human must be named.
III. Key Case Laws Relevant to AI-CDSS
1) Mayo Collaborative Services v. Prometheus Laboratories (2012)
Facts:
Prometheus claimed a method for optimizing drug dosage by measuring metabolite levels in a patient’s blood.
The Supreme Court examined whether this method was patentable.
Holding:
The Court ruled that the method claimed only a natural correlation (metabolite levels and drug efficacy/toxicity) and routine steps.
Such claims are not patentable because they are directed to a natural law.
Implication for AI-CDSS:
AI methods that just analyze correlations in medical data without adding a technological improvement may not be patentable.
2) Ariosa Diagnostics v. Sequenom (2015)
Facts:
Sequenom developed a method to detect fetal DNA in maternal blood (non-invasive prenatal testing).
Ariosa challenged the patent, arguing it was invalid.
Holding:
The court invalidated the patent under the Mayo principle: it involved natural phenomena (fetal DNA in maternal blood) and conventional lab techniques.
Implication:
AI-CDSS systems that automate existing diagnostic methods may be vulnerable to invalidation if they rely solely on natural correlations.
3) Enfish, LLC v. Microsoft (2016)
Facts:
Enfish claimed a self-referential database software that improved how computers store and retrieve data.
Holding:
The court held the claims patentable because they improved computer functionality, not just an abstract idea.
Implication for AI-CDSS:
AI systems that improve computational methods, data handling, or model efficiency are more likely to be patentable.
4) Diamond v. Diehr (1981)
Facts:
Diehr patented a process for curing rubber using a mathematical formula to determine the optimal temperature.
Holding:
The Supreme Court held the patent valid because the invention involved a physical process incorporating mathematics, not just an abstract formula.
Implication for AI-CDSS:
Integration of AI into a concrete medical workflow or machine process (like automated infusion pumps or robotic surgery) can support patent eligibility.
5) Thaler v. Comptroller-General of Patents (DABUS Cases)
Facts:
Stephen Thaler tried to patent inventions where an AI system (DABUS) was listed as the sole inventor.
Holding:
Courts in the US, UK, and Europe ruled AI cannot be an inventor; only a human can be legally recognized.
Implication for AI-CDSS:
Even if AI creates a novel diagnostic method, patents must list human inventors.
6) IBM Watson Health v. MD Anderson
Facts:
IBM’s Watson AI system was deployed in oncology to assist in clinical decisions.
MD Anderson alleged IBM failed to disclose patentable improvements and questioned ownership of inventions created during collaboration.
Outcome:
The dispute was settled with clear IP ownership and licensing agreements.
Implication:
Collaborative AI-CDSS projects must have explicit agreements on patent rights and ownership, otherwise disputes are likely.
IV. Key Takeaways for AI-CDSS Intellectual Property
| Issue | Takeaways |
|---|---|
| Patent Eligibility | Must include technical innovation, not just data analysis or natural correlations (Mayo, Ariosa). |
| Software Improvement | Claims improving computing methods, speed, or workflow are stronger (Enfish, Diamond v. Diehr). |
| Inventorship | Only humans can be inventors (Thaler v. Comptroller). |
| Collaborative Projects | Clearly define IP ownership in agreements (IBM Watson v. MD Anderson). |
| International Considerations | Must satisfy novelty, inventive step, and technical contribution (similar principles apply globally). |
RELATED Blog
comments