Patenting Artificial Intelligence: Issues of Obviousness, Inventorship, and Patent Eligibility

• Jurisdiction: United States,Federal
• Publication year: 2018
• Citation: Vol. 1 No. 5

Susan Y. Tull and Paula E. Miller^*

Patent protection must keep pace with the growing applications of artificial intelligence in medical and pharmaceutical technologies. The rise of "thinking machines" raises questions regarding the application of personhood to patent law, including the definition of a "person" of skill in the art, predictability, inventorship, and subject matter eligibility. This article will address these questions in light of the recent technological advances.

Artificial intelligence ("AI") is rapidly transforming the world of medicine, as the recent decades have marked a surge in the development of medical AI.¹ These thinking machines are now used in diagnosis, treatment, and drug development. As the technology advances, so too must our understanding of patent law and patent protection. The use of AI in these fields raises several issues, all hinging on the question of personhood and human contributions, affecting both inventorship (and ownership) and patentability (including subject matter eligibility and predictability). This article addresses these questions in turn after addressing the recent advances in medical AI.

Artificial Intelligence in Medicine

AI techniques utilized in medicine include artificial neural networks, fuzzy expert systems, evolutionary computation, and hybrid intelligent systems.²

Artificial neural networks are used extensively in clinical diagnosis and image analysis because of the parallel processing power that allows the networks to learn from historical examples and known patterns.³ Artificial neural networks have been used for diagnosing prostates as benign or malignant, cervical screening, and imaging analysis (including radiographs, ultrasounds, CTs,

[Page313]

and MRIs), as well as for analyzing heart waveforms to diagnose conditions such as atrial fibrillation and ventricular arrhythmias.⁴

For example, researchers at Stanford University trained a deep convolutional neural network to classify skin lesions into either benign or malignant groupings based on known images, using only pixels and disease labels as inputs.⁵ The researchers started with an algorithm developed by Google to perform image recognition⁶ and then trained their neural network to recognize skin cancer using 129,450 clinical images of 2,032 different diseases.⁷ The neural network was then tested against board-certified dermatologists on clinical images that had been confirmed through biopsy.⁸ The AI performed on par with the certified dermatologists, demonstrating that the AI was capable of classifying skin cancer with the same level of competence as the trained dermatologists.⁹

As yet another example, medical chatbots utilize neural networks to learn from medical textbooks, scientific research, patient records, and messages between actual patients and doctors.¹⁰ The AI chatbot is constantly learning and can be kept up to date on the latest medical research.¹¹ Baidu, a Chinese search engine, utilizes a chatbot named Melody within its Baidu Doctor app.¹² When a patient asks a question to the doctor, the chatbot asks appropriate follow-up questions to help learn more about the patient's symptoms so the doctor can make a more informed decision on treat-ment.¹³ Interventional radiologists at the University of California at Los Angeles have developed a chatbot to assist physicians in providing real-time evidence-based answers to the patient about the next phase of treatment, or information about their interventional radiology treatment.¹⁴

Fuzzy logic AI is applicable in medicine because diseases, symptoms, and diagnoses are described in imprecise and terms.¹⁵Because fuzzy logic rests on the premise that everything is a matter of degree, it can recognize "partial truth logics," beyond just the true and false values applied in traditional programming.¹⁶ Fuzzy logic AI has been applied to cancer diagnosis for lung cancer, acute leukemia, breast cancer, and pancreatic cancer.¹⁷ Fuzzy logic has been applied to diagnosis of other conditions, including tuberculosis, aphasia, arthritis, and hypothyroidism.¹⁸

"Evolutionary computation is the general term for several computational techniques based on natural evolution process that imitates the mechanism of natural selection and survival of the fittest in solving real-world problems."¹⁹ Genetic algorithms are the

[Page314]

most widely used form of evolutionary computation in medicine, creating numerous solutions to a single problem, and then evolving those solutions from one generation to the next to arrive at the best solution.²⁰ Evolutionary computation has been used in diagnosis, prognosis, imaging, and signal processing.²¹

Combining these AI techniques generates hybrid intelligent systems that incorporate the advantages of each technology.²² For example, the combination of neural networks and fuzzy logic or "neuro-fuzzy" systems have become popular because they can absorb some of the "noise" generally present in the neural network.²³

Uncertainties in Patenting AI—Subject Matter Eligibility

As the use of AI in medicine becomes ever more prevalent, the patent system must answer increasingly difficult questions regarding the protection afforded these technologies. Perhaps the most significant question is that of subject matter eligibility. With the Supreme Court decisions in Alice and Mayo, the hurdle to meet subject matter eligibility has grown ever higher.²⁴

Subject matter eligibility is one of the core criteria for receiving a patent, in addition to novelty and nonobviousness. An invention must contain patent-eligible subject matter in order to receive patent protection. 35 U.S.C. § 101 states that "[w]hoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor." Abstract ideas, laws of nature, and natural phenomena are excluded from patentable subject matter.²⁵The U.S. Supreme Court has further enunciated the requirement for subject matter eligibility, ultimately laying out a two-part test that must be met by any claimed invention.

In Mayo, the Supreme Court invalidated issued patent claims directed to the relationship between the concentrations of certain metabolites in the blood and the likelihood that a drug dosage would prove ineffective or cause harm for failing to meet this requirement.²⁶ The Supreme Court held that the claims were not subject matter eligible under 35 U.S.C. § 101 because the claims provided "instructions [that] add nothing specific to the laws of nature other than what is well-understood, routine, conventional activity, previously engaged in by those in the field."²⁷ According to

[Page315]

the Court, the dosage limits at which a drug would prove ineffective or cause harm was a law of nature that was unpatentable, and the claims merely instructed doctors to apply this law of nature using techniques that were already known.

Alice addressed the holding in Mayo, further enunciating a two-step test for subject matter patent eligibility: (1) determine whether the claims are directed to a patent-ineligible concept (laws of nature, abstract ideas, and natural phenomena); and (2) determine whether the claim's elements, considered both individually and as an ordered combination, transform the nature of the claims into a patent-eligible application.²⁸ If a claim is directed to a patent-ineligible concept and the claim's elements do not transform the nature of the claim, then it will fail to meet § 101.

These two Supreme Court cases present a hurdle that medical AI inventions will have to overcome in order to receive patent protection. Current AI medical device/system patents can be directed to both the methods and apparatuses that perform the above-described analyses. Many AI medical patents are directed to the AI algorithms and the machines used to generate those algorithms.²⁹ As described above, AI has been found to be extremely successful in diagnosis and prognosis, relating known images to new cases and extrapolating based on the similarities or differences between the two. In some instances, this is the same process followed by a doctor or medical expert, just with greater efficiency or accuracy. The steps for diagnosis struck down in Mayo echo the steps taken in many medical AI algorithms. Practitioners and inventors alike will need to carefully consider the full scope of eligible subject matter in order to ensure that a patent can be obtained from the U.S. Patent and Trademark Office ("PTO") and maintained through any subsequent challenges.

Indeed, the Federal Circuit has already found revolutionary diagnostic technology to be patent-ineligible subject matter under the Mayo/Alice framework. In Ariosa Diagnostics, Inc. v. Sequenom, Inc., the court concluded that a novel method of prenatal diagnosis of fetal DNA was not directed to patent-eligible subject matter, despite agreeing that the claimed method "reflects a significant human contribution . . . that revolutionized prenatal care."³⁰ The patent claims were generally directed to detecting the presence of cell-free fetal DNA in maternal plasma. Because the presence of cell-free fetal DNA was a natural phenomenon, the court turned to the second step in the Mayo/Alice framework—whether the claim

[Page316]

contained an inventive concept sufficient to transform the naturally occurring phenomenon into patent-eligible subject matter.³¹ The court found that the second step was not met because the method steps "were well-understood, conventional, and routine," despite acknowledging their breakthrough nature.³²

More recently, the Federal Circuit found methods for detecting myeloperoxidase ("MPO") in blood, and correlating the results to cardiovascular risk, to be directed to patent-ineligible subject matter in Cleveland Clinic Foundation v. True Health Diagnostics LLC.³³ Although Cleveland Clinic argued that the discovery of the correlation was groundbreaking, the Federal...