In a series of decisions beginning in 2010, the US Supreme Court reshaped the law of patentability. The changes left the validity of thousands of issued patents in doubt and created deep and wide uncertainty and confusion about what subject matter is actually patentable. Considerable uncertainty remains and the law is still in flux, but recent decisions of the US Court of Appeals for the Federal Circuit have begun to provide principles governing whether a particular invention is patentable.

The basics of patentability

Any discussion of what can be patented begins with Section 101 of Title 35 of the US Code, which establishes the categories of patentable subject matter. Patents may issue for any new and useful “process, machine, manufacture, or composition of matter”. The courts have described these four statutory categories as follows:

  • process – an act or series of acts or steps;
  • machine – a concrete thing consisting of parts or certain devices and combinations of devices;
  • manufacture – an article produced from raw or prepared materials by giving to these materials new forms, qualities, properties or combinations, whether by hand labour or by machinery; and
  • composition of matter – all compositions of two or more substances and all composite articles, whether they be the results of chemical union or of mechanical mixture, or whether they be gases, fluids, powders or solids, for example.

These categories encompass many kinds of invention, although they do not extend to every product of human ingenuity. For example, it has been held that a transitory propagating signal as such may not be patented, precisely because it is not a process, machine, manufacture or composition of matter. Contracts and corporations are unpatentable for the same reason.

But the four categories are merely the beginning of the analysis. Courts have created judicial exceptions to patentability, applicable to inventions that fit within an otherwise patentable category. Specifically, a patent may not be granted on:

  • laws of nature;
  • natural phenomena; and
  • abstract ideas.

The stated goal of these exceptions is to prevent improper tying-up of the building blocks of human ingenuity.

Supreme Court decisions

In a series of decisions beginning in 2010, the Supreme Court considered the judicial exceptions and transformed the law of patentability. These cases were as follows:

  • Bilski v Kappos (561 US 593) (2010) – rejected the test created by the Federal Circuit to determine the patentability of methods. The rejected rule was that a method is unpatentably abstract unless it is tied to a particular machine or effects a particular transformation of matter. The Supreme Court held that this rule, while a useful guideline, could not be the only criterion for patentabilty; but the court did not say what the additional criteria were.
  • Mayo Collaborative Servs v Prometheus Labs, Inc (132 S Ct 1289) (2012) – invalidated a patent directed to a medical diagnostic method. The claimed method involved administering a specified drug to a patient, measuring the amount of the drug in the patient’s blood and then adjusting the next dose based on the measurement, according to criteria recited in the claims. The court found that the claims covered only natural phenomena and all applications of those phenomena, so the patent was impermissibly directed to a natural law as such.
  • Assoc for Molecular Pathology v Myriad Genetics, Inc (133 S Ct 2107) (2013) – invalidated a patent directed to DNA found in nature. The patent owner had found and isolated genes associated with an increased risk of breast cancer. The court held that the naturally occurring DNA was a product of nature and therefore unpatentable, despite its discovery by the inventors.
  • Alice Corp v CLS Bank Int’l (134 S Ct 2347) (2014) – invalidated a software-implemented business method as directed to an abstract idea, but without defining ‘abstract idea’. The court characterised the claims as directed to the abstract financial concept of intermediated settlement, which was an unpatentable abstract idea. The court further held that an unpatentable abstract idea did not become patentable merely because it had been claimed in the form of a computerised implementation of that idea.

A result of these decisions was the Mayo-Alice framework for judging patentability, which requires a two-part analysis. Step 1 is the threshold inquiry, asking whether the claimed invention is within the four statutory categories. If not, the invention is unpatentable and the analysis ends.

Step 2 applies if the invention is within one or more of the statutory categories, and it has two sub-parts. Step 2A is the determination of whether the claimed invention falls within the judicial exceptions to patentability; if none of the exceptions applies, the subject matter is patentable. If an exception does apply, Step 2B is the determination of whether the patent claims include an “inventive concept” that is “significantly more” than the exception.

All four decisions were disruptive, but none more so than Alice, which threw into question the validity of thousands of patents that, following common practice, had claimed business methods in terms of a computer system that is programmed to carry them out. It further threatened the patentability of most software-defined inventions, in part because of the nature of software development. In software development, there is often little distance between an idea and its realisation in executable code. Indeed, it is common in the industry for engineers to communicate ideas in the form of working source code.

A corollary, though, is that a developer often needs to know only what a computer is intended to do to be able to write a program to do it. Patents directed to software-implemented inventions commonly describe and claim the invention in terms of the inventive idea and this disclosure suffices to enable persons skilled in the computer arts to practise the idea. However, as a consequence, even patents directed to specific, concrete inventions often describe and claim the invention at a more abstract level than is common in other fields. Thus, Alice left software-implemented inventions particularly vulnerable.

Alice had this effect even though, read out of context, it can be taken as a relatively narrow decision. The unpatentability of abstract ideas was well established at the time, as was the doctrine that an unpatentable abstract idea does not become patentable simply because it is claimed in a form that amounts to: “Apply the abstract idea.” One natural reading of Alice is that it holds merely that an unpatentable abstract idea does not become patentable simply because it is claimed in a form that amounts to: “Apply the abstract idea using a computer.”

A big part of the problem was that the Supreme Court refused to say what makes an idea unpatentably abstract or to define ‘inventive concept’. This refusal was particularly striking because of the court’s recognition that “[a]t some level, all inventions… embody, use, reflect, rest upon, or apply laws of nature, natural phenomena, or abstract ideas”. If all inventions are at some level abstract, there is a clear need to find the level beyond which an invention is too abstract, but the court did not begin to address this.

The Federal Circuit after Alice

A great deal of uncertainty followed Alice. Some suggested that all inventions expressed solely in software would be unpatentable going forward and many recognised that previous drafting techniques left many existing patents and pending patent applications vulnerable to invalidation, even if they were directed to theoretically patentable subject matter.

The US Court of Appeals for the Federal Circuit is the intermediate appellate court with sole jurisdiction over patent matters. In the aftermath of Alice, the Federal Circuit was expected to be the main source of guidance on what Alice meant. But the Federal Circuit had no more basis to proceed than anyone else and the first opinions in this area show the court’s groping for applicable principles.

An important early case was DDR Holdings, LLC v (773 F3d 1245 Fed Cir 2014). DDR was important because it was the first post-Alice case in which the Federal Circuit sustained claims to a software-implemented invention. The court did not clearly express its decision in terms of the Mayo-Alice framework; that is to say, it was not clear whether:

  • the claimed invention was not directed to an abstract idea; or
  • the claims, although directed to an abstract idea, included a sufficient inventive concept.

Later cases went further and held that software-implemented inventions are not per se directed to abstract ideas, which means that Step 2B of the Mayo-Alice analysis is not always needed for software-related inventions.

Since then, several principles have emerged from the Federal Circuit’s decisions:

  • Improvements to computing and relating technologies are strongly likely to be patentable.
    • Following a suggestion in Alice, the Federal Circuit held in Enfish, LLC v Microsoft Corp that a claim directed to a specific improvement in the way that computers operate was not directed to an abstract idea. Permissible improvements include improvements effectuated purely through software running on generic computing hardware; for example, in Enfish, the claimed invention was an improvement to the way databases organise information (822 F3d 1327 Fed Cir 2016).
  • Mere computerisation of existing business practices is strongly likely to be unpatentable.
    • This was the basis for the results in Bilski and Alice. In Bilski the Supreme Court characterised the claims as directed to the concept of hedging risk, as such, and in Alice the court said that the claims were directed to intermediated settlement. In both cases, the court found that these were well-known, long-established business practices that did not become patentable simply because they had been implemented by software executing on generic computing hardware.
    • A method that can be carried out solely in the human mind, or with pen and paper, will be considered abstract. If a method is to be performed by a computer system, the claims should say so explicitly.
  • Inventions in which the main novelty relates to some combination of human relationships, including financial or legal obligations, are strongly likely to be unpatentable.
    • As a practical matter, this kind of invention has usually been treated as mere computerisation of existing business practices, even if the patent office or the court must disregard aspects of the claimed invention in doing so. This sometimes takes the form of characterising the invention as merely applying well-known techniques in a new commercial or technological environment.
    • When the asserted novelty lies in setting prices, for example, the Federal Circuit has found the claims to be directed to abstract ideas as such; without an additional inventive concept, the invention has been found to be unpatentable.
  • In contrast, automation of earlier practice may be patentable when the inventive aspect lies in the solution to a problem that arises because of automation.
    • DDR Holdings involved generating composite e-commerce websites in which content from third parties was given an appearance consistent with the vendor’s own products. The Federal Circuit held that giving all pages a consistent look and feel was a solution “necessarily rooted in computer technology to overcome a problem specifically arising in the realm of computer networks”.
    • Similarly, the Federal Circuit has found that the automation of some aspects of computer animation is patentable. The patent was not directed simply to automating tasks that previously had been done by human animators, but included replacement of the animators’ case-by-case judgement with specific new rules that the computer would follow.
    • Graphical user interfaces (GUIs) have been held to be patentable where they include new elements or combinations of elements, even when the problem they solve came out of the kind of business. For example, an interface was found to be patentable because it brought greater speed, accuracy and usability to computerised trading. In contrast, a GUI will not be patentable if it uses only conventional elements in conventional ways.
    • This is not the same as merely describing or claiming an invention at a greater level of specificity. For example, if a claim is directed to the abstract idea of searching in an index, the flaw is not fixed by specifying further that the entries in the index are XML tags, unless the XML elements are used in a way that is not itself conventional.
  • A patentable inventive concept can lie in the non-conventional and non-generic use of known, conventional pieces.
    • In an example case, the Federal Circuit sustained a patent directed to web filtering, even though filtering content was a well-known and longstanding way to organise human behaviour. The court found an inventive concept in the claimed improvement over existing filtering systems.
  • To make the strongest case for patentability, a method should be expressed, to the greatest extent possible, in terms of the actions that a computer system performs, rather than the results of those actions.
    • The Federal Circuit stated this principle in considering the computer animation case above. One reason for the court’s decision in that case was its finding that the claims incorporated specific rules for automating the steps of animation and were not directed to automating computer animation as such.

It is important not to overstate the applicability of these principles. The law is still in flux and the above principles must be considered at least somewhat tentative. Further – and particularly relevant to practitioners – they are still filtering down to the federal trial courts and the US Patent and Trademark Office (USPTO). For example, the USPTO’s official guidance directs patent examiners to resolve questions of subject matter patentability by comparing the claims before them to the specifics of the claims at issue in published Federal Circuit opinions.

However, principles are indeed emerging and they show the beginnings of the path to patentability for software-implemented inventions.

This article first appeared in IAM. For further information please visit