Panel Criticizes Bilski Machine-or-Transformation Test
By Kathryn Freund – Edited by Ryan Ward
On Thursday, April 1st, JOLT hosted a panel discussion at Harvard Law School on the impact of In Re Bilski, a 2008 en banc decision by the Court of Appeals for the Federal Circuit holding that the machine-or-transformation test is the proper method for determining patent-eligible subject matter. Digest previously covered the Bilski decision here. The case was appealed to the Supreme Court, which heard oral arguments in November 2009.
The three panelists, Don Steinberg, Professor Kevin Collins of the University of Indiana Maurer School of Law, and Professor Michael Meurur of Boston University School of Law, examined the Bilski holding from different litigation and economic perspectives, and wagered predictions on the forthcoming Supreme Court decision. The panelists agreed that the Bilski machine-or-transformation test rightly excludes software and business method patents, but fails as a useful test. They went on to agree that the Supreme Court will likely strike down the Bilski patent, but is unlikely to provide a concrete new test for courts to follow.
Don Steinberg, the chair of the Intellectual Property Department at WilmerHale and a JOLT alumnus, opened by describing the Bilski machine-or-transformation test. The two-part test holds that something is patentable if it is (1) “tied to a particular machine or apparatus,” or (2) “transforms a particular article into a different state or thing.”
Steinberg criticized the Bilski test as a “simple sounding” but “impossible to actually apply.” While the machine-or-transformation test “gets 80% of the way there” and eliminates patent eligibility for general principles [and] algorithms, the “Supreme Court cases on which the court relied seemed to say that the machine-or-transformation test [was] good guidance, not the exclusive test — and the test doesn’t work well in a lot of cases.” Steinberg specifically argued that the machine portion of the test does not capture “nonconventional uses” of existing machines. Thus, he claimed that patent eligibility cannot focus solely on the machine-or-transformation test, but instead should look for innovation in the product or method.
Referring to patents more generally, Steinberg criticized over breadth, in which patentees try to claim “more than they invented or contributed.” For example, Steinberg argued that, in Assn. for Molecular Pathology v. USPTO, Myriad Genetics attempted “to patent the concept” by claiming a method without contributing anything not already existing in nature. The Myriad court held that simply describing a claim as a method or system does not make the claim patentable. Steinberg suggested that this holding “should not be limited to method claims.”
Professor Kevin Collins, a Professor of Law at Indiana University whose research focuses on patent law, described Bilski as a decision which uses “intangibility of an invention as a proxy for what should not be eligible for patent protection.” Professor Collins described the main benefit of the machine-or-transformation test as excluding business methods and software, areas where patents impose significant economic costs, from patent protection.
Professor Collins explained that the tangibility portion of the test in Bilski originated from the “determine and infer” inquiry used in Prometheus Labs., Inc. v. Mayo Collaborative Servs. He rejected the tangibility proxy, saying, “[it] seems really dumb to use intangibility as a filter to determine which claims are patent eligible.” Courts require that the tangible physical transformation be “central to the purpose of the claimed method” in the determining step of the “determine and infer” inquiry. Collins acknowledged that this requirement could be a public proxy for how far upstream in the research process “determine and infer” can lead to patentability. However, he argued that a knowledge/embodiment dichotomy would provide a better test than tangibility to determine patentability.
Both Steinberg and Professor Collins proposed applying the copyright idea/expression dichotomy to determine patent subject matter eligibility. Steinberg applied this concept to patent law, analogizing an algorithm to an idea, which should not be patentable, and comparing a process involving an algorithm to an expression, which should be patentable.
The final speaker, Professor Michael Meurur, a Professor of Law at Boston University, argued for a “categorical exclusion” of business method patents because these inventions do not lend themselves to property-based incentives. Professor Meurur offered empirical evidence to demonstrate that eliminating business method patents, which do not provide clear property boundary lines, will decrease litigation and “reduce the tax on innovation that, in aggregate or in average, business method patents are creating.” For example, financial method patents were found to be thirty-nine times more likely to be asserted in a lawsuit than other types of patents.
The panelists agreed that the Supreme Court will likely uphold the Federal Circuit’s rejection of Bilski’s patent. Steinberg predicted that the Court will go back to the Benson-Flook-Diehr trio of patent cases without providing much guidance or an exclusive test for dealing with business method claims. Professor Collins agreed and expects many concurring opinions in the decision. Professor Meurur suggested that the Court might look to the industrial arts test used in Europe as an alternative test for determining patent eligibility.
Kathryn Freund is a 1L at Harvard Law School.