Legal Certainty Challenges In Algorithmic Incremental Innovation.
1. Introduction: Legal Certainty in Algorithmic Incremental Innovation
Algorithmic incremental innovation refers to situations where an existing algorithm is improved step by step rather than through radical invention. While this approach accelerates technological progress, it creates legal uncertainty in patent law because:
Small modifications may or may not be patentable.
Determining inventive step (non-obviousness) is challenging.
The prior art landscape changes rapidly due to algorithmic improvements.
Ownership and inventorship questions arise when AI contributes to incremental improvements.
Legal certainty requires clarity on what is patentable, who owns the invention, and how prior incremental innovations are treated.
2. Key Case Laws Illustrating Legal Challenges
Case 1: Alice Corp. v. CLS Bank International (2014, US Supreme Court)
Facts:
Alice Corp. had patents on a computerized scheme for mitigating settlement risk in financial transactions.
The innovation was algorithmic and mostly incremental improvements over existing financial systems.
Legal Issue:
Whether incremental improvements in software and algorithms constitute a patentable “abstract idea” under §101 of the U.S. Patent Act.
Holding:
The Supreme Court ruled that implementing an abstract idea on a computer does not automatically make it patentable.
Incremental improvements must involve an “inventive concept” beyond routine computer implementation.
Impact:
Reinforced legal uncertainty in algorithmic incremental innovation.
Small tweaks or optimizations to existing algorithms might not meet patentability requirements, causing companies to hesitate in patenting incremental innovations.
Case 2: Enfish, LLC v. Microsoft Corp. (2016, US Federal Circuit)
Facts:
Enfish claimed a patent for an improved self-referential table in a database.
Microsoft argued that the patent was abstract and therefore unpatentable.
Legal Issue:
Are incremental improvements to database algorithms patent-eligible?
Holding:
The Federal Circuit held that a specific improvement to the functioning of a computer can be patentable.
Unlike Alice, Enfish emphasized that if the improvement solves a technical problem in a novel way, incremental innovations can be protected.
Impact:
Provided a legal pathway for algorithmic incremental innovation.
Introduced the idea that technical improvements in algorithms can mitigate uncertainty around patent eligibility.
Case 3: T 1227/05 – EPO (Two-Tiered Algorithms in Simulation Software)
Facts:
The European Patent Office (EPO) considered a patent application for incremental improvements in simulation software algorithms.
Legal Issue:
Whether improvements in existing algorithms without a new technical effect are patentable.
Holding:
The Board of Appeal ruled that incremental changes must demonstrate a technical effect beyond mere mathematical optimization.
General improvements in efficiency alone may not suffice.
Impact:
Created legal uncertainty for developers of algorithmic incremental innovations in Europe.
Reinforced that technical contribution is the key determinant in algorithmic patents.
Case 4: Apple Inc. v. Motorola Inc. (2012, US District Court, W.D. Wis.)
Facts:
Apple sued Motorola for patent infringement involving algorithms in mobile devices.
Many patents involved small improvements over existing software.
Legal Issue:
How should incremental algorithmic improvements be evaluated for infringement and validity?
Holding:
The court emphasized that obviousness (35 U.S.C §103) must be carefully applied in incremental algorithmic improvements.
Incremental innovation that seems minor could still be patentable if it solves a non-obvious technical problem.
Impact:
Showed that incremental innovations create uncertainty for both patentees and infringers.
Even minor algorithmic changes can trigger litigation if technical impact is not clearly documented.
Case 5: T 0930/18 – EPO (Machine Learning Optimization Algorithm)
Facts:
Applicant sought patent for a minor modification of a machine learning algorithm improving training speed.
Legal Issue:
Is incremental improvement in training speed patentable under the EPO?
Holding:
The EPO emphasized the “technical character” of improvements.
Purely mathematical or statistical improvements without technical application were rejected.
Demonstrating real-world implementation and effect was crucial.
Impact:
Incremental improvements in AI/ML algorithms face uncertainty unless technical applicability is clearly shown.
Shows that minor algorithmic gains in efficiency may not guarantee patent protection.
Case 6: Thales Visionix Inc. v. United States (2017, Court of Federal Claims)
Facts:
The patent involved incremental improvements to sensor-fusion algorithms for navigation systems.
Legal Issue:
Whether iterative improvements in algorithmic sensor integration constitute patentable subject matter.
Holding:
The court recognized the technical contribution of the incremental improvements, emphasizing that they solved a specific technical problem in the physical system.
Impact:
Reinforced that incremental algorithmic innovations tied to physical applications enjoy stronger legal certainty.
Highlights the EPO vs. US approach: technical application increases certainty.
3. Patterns and Challenges Observed
Abstract Idea Risk: Courts often struggle to separate algorithmic ideas from technical application. Incremental improvements in abstract algorithms may not be patentable (Alice, EPO cases).
Technical Effect Requirement: Incremental improvements must solve a technical problem in a non-obvious way (Enfish, T 0930/18).
Ownership Ambiguities: When multiple parties or AI systems contribute to incremental innovation, inventorship disputes can arise (noted in AI patent jurisprudence in US & EU).
Rapid Obsolescence: Algorithmic improvements can become prior art very quickly, creating uncertainty in novelty assessment.
EPO vs US Approaches:
EPO: Focus on technical character and real-world effect.
US: Focus on inventive concept and non-obvious technical improvement.
4. Summary Table
| Case | Jurisdiction | Incremental Innovation Issue | Key Principle |
|---|---|---|---|
| Alice v. CLS | US | Abstract financial algorithms | Abstract ideas not patentable; need inventive concept |
| Enfish v. Microsoft | US | Database algorithm optimization | Technical improvement can be patentable |
| T 1227/05 | EPO | Simulation software algorithm | Must demonstrate technical effect |
| Apple v. Motorola | US | Mobile algorithm improvements | Minor improvements can be patentable if non-obvious |
| T 0930/18 | EPO | ML algorithm optimization | Implementation and technical application required |
| Thales Visionix | US | Sensor fusion algorithm | Technical solution to physical problem strengthens patentability |
✅ Conclusion
Legal certainty in algorithmic incremental innovation is fragile because:
Minor algorithmic tweaks can be abstract or obvious.
Courts emphasize technical contribution for patentability.
Evolving AI/ML algorithms complicate prior art analysis and inventive step evaluation.
Practical takeaways for innovators:
Document technical contribution and real-world effect of each incremental change.
Consider patenting broader systems rather than tiny algorithmic tweaks.
Monitor evolving case law in US and EU; the landscape is dynamic.
RELATED Blog
comments