PATENT UNCERTAINTY: TOWARD A FRAMEWORK WITH APPLICATIONS

idea of testing an equipment operator for impairments using an unspecified ‘expert system’ running on equipment that already exists in various vehicles.”). 2016] PATENT UNCERTAINTY 1141 the patent system to develop such a technology, and seek to make a profit by promoting the technology to car-makers or to regulatory authorities. In addition, a car manufacturer, if it had a sufficient incentive to develop such a technology, would use it to gain an advantage in the market over rivals. The individual inventor, in contrast, has no incentive to distort the market in favor of one particular firm; his incentive is to license the innovation to all. These arguments suggest that society should be reticent to adopt rules in the patent system that deny patent law’s encouragement to certain safety-enhancing innovations. Important distinctions between web navigation patents and impairment testing should be noted. Web navigation is one of many functions that every business has an incentive to optimize for its customers, whether patents are available or not. Competition will drive firms to adopt such functions. General technologies that improve safety across an entire industry are not necessarily functions that every business has an incentive to develop, a proposition established in the law since Learned Hand’s opinion in The T.J. Hooper.105 Some of the victims of unsafe conditions may not be customers of the industry—and the industry will have little incentive to devote resources to reduce such external losses.106 The costs of developing a safety technology may be high, the benefits may not be appropriable, and competition may not reward such innovations. This suggests at least one principle that should be adopted for conceptualfunctions patents: Conceptual-functions patents related to optimizing consumer-firm interactions should be presumptively ineligible for patents. Every business has an incentive to optimize its interactions with consumers— to make its websites navigable by mobile devices, to order its messages to and from consumers efficiently,107 to suggest alternative and higher priced goods or 105 60 F.2d 737, 740 (2d Cir. 1932) (recognizing that though many boats used radio technology, it had not yet become an industry standard and the court was in no position to demand that the rest of the industry conform). 106 Richard A. Posner, A Theory of Negligence, 1 J. LEGAL STUD. 29, 39 (1972) (“Suppose the only benefit of a safety appliance is to a stranger to the industry in our earlier sense—someone with whom the enterprise has no contractual relationship and will not enter into one because of transaction costs. No firm in the industry will have an incentive to install the appliance.”). 107 Two business-messaging patents asserted against Hewlett-Packard and Adobe were recently found invalid under Alice. YYZ, LLC v. Hewlett-Packard Co., Civ. No. 13-136SLR, 2015 WL 5886176, at *9 (D. Del. Oct. 8, 2015) (holding that the custom messaging component was neither specifically claimed nor sufficiently innovative and was therefore not patent-eligible). In a different and recent case, Home Depot was sued for infringing a patent that lets customers choose whether to have their receipts printed or emailed to them. Complaint at 3-4, eCeipt LLC v. Home Depot, Inc., No. 2:15-cv-01672 (E.D. Tex. filed Oct. 21, 2015). Based on the argument of this paper, this claim should be rejected under Alice. 1142 BOSTON UNIVERSITY LAW REVIEW [Vol. 96:1117 services,108 and so on. The patent system should not grant exclusive rights in customer-oriented or supplier-oriented business functions. If the presumption against patentability suggested by Alice is to be extended beyond algorithmic patenting into conceptual-functions patenting, which I have suggested may be desirable, the next set of concepts to bring under Alice are those involving the customer-firm or supplier-firm interface. The doctrinal test of Alice can remain without doing any harm; it is too vague to have much of an impact anyway on its own. If there is any practical rule that Alice should stand for, it is for a presumption against patent eligibility for both algorithmic software applications and for business methods that enable firms to sell, advertise, or interact with consumers more efficiently. This subset of conceptual functions should be left to the process of Schumpeterian competition.109 For conceptual functions with potentially life-saving applications (such as the driver impairment testing claimed by the Roe patent), Alice must tread much more cautiously.110 The reason is that the additional incentive provided by the patent system for life-saving innovations should be considered worth preserving, and the market may be insufficient as a spur to such innovation. On the other hand, these patents have to be assigned boundaries. Although the principle disfavoring abstraction is well entrenched in patent law, it should be modified to constrain patentees more effectively in this area. A doctrine shifting the burden of proof to the patentee to show that his patent really does have boundaries and that there are realistic methods of innovation within the same conceptual function that are not preempted may be appropriate. Alice may have had this effect already. 108 See Tuxis Techs., LLC v. Amazon.com, Inc., No. 13-1771, 2014 WL 4382446 (D. Del. Sept. 3, 2014) (invalidating under Alice a patent for “upselling” (i.e., suggesting pricier options) to online consumers). 109 CASS & HYLTON, supra note 8, at 71 (“Schumpeterian creative destruction[] involves the continual introduction of new methods that lead at times to temporary monopolies but are eventually copied by competitors. Patents might provide an additional incentive to develop new methods, but they will also obstruct the process of dissemination and emulation that is core to dynamic competition.”). 110 I should be clear that this suggestion is limited to Alice as a doctrine for invalidating patents. More traditional doctrines for invalidating patents, such as obviousness, do not have to tread cautiously as suggested here. Traditional grounds for invalidating patents have been in operation for many years and should not be affected by Alice. For example, the Patent Trial and Appeal Board granted TRW Automotive US LLC’s petition for inter partes review of a driver-assistance camera patent belonging to Magna Electronics Inc. after TRW demonstrated a reasonable likelihood it would be able to prove the patent was anticipated by previous technology. TRW Auto. US LLC v. Magna Elec. Inc., No. IPR2015-00923, 2015 WL 5818392 (P.T.A.B. Oct. 1, 2015). Although this is an example of a “driver safety” innovation similar to the Roe patent discussed in the text, the difference is that the PTAB reexamined the patent on obviousness grounds, not on the basis of Alice. 2016] PATENT UNCERTAINTY 1143 My suggestion goes beyond life-saving innovations. Inventions occur along a spectrum from fundamental science, to production-oriented engineering, to methods that serve consumers or end-users efficiently. Fundamental science, such as math and physics, has in large parts been and should be considered ineligible for patents. Innovations that make the firm-consumer or firm-supplier interface more efficient should also be ineligible because the market is a sufficient spur. The production-oriented engineering level is where patents are most effective in enhancing social welfare, and much of this innovation is taking the form of software today. Alice is clearly applicable to this type of innovation. Where there is a danger of granting a property right in the abstract conceptual function, the burden should shift to the patentee to argue that his right will not preempt future innovation within the same conceptual function during the patent’s term. The precise boundary of Alice is unsurprisingly impossible to delineate with precision on the basis of general principles. Software that primarily optimizes the consumer-firm or supplier-firm interaction, or similarly facilitates business processes, should face a presumption of ineligibility under Alice. Potentially life-saving technology generally should not face such a presumption. Software in the online security area seems to fall somewhere in between these two poles.111 Online security software may serve to protect a firm’s data from expropriation from rivals, which is a purely business-centered function that is likely to be spurred sufficiently by market forces without the need for patent protection. Some online security, by contrast, may serve to protect consumers from identity theft, which provides benefits beyond the boundaries of the firm’s own balance sheet. Indeed, given that the primary costs of identity theft fall on the consumer rather than the firm, a profit-oriented firm may have weak incentives to protect the consumer from such theft. The firm may trade off greater convenience in consumer transactions in exchange for less protection of the consumer. Just as in the case of the Lincoln patent, the market may provide inadequate incentives for firms to invest in innovations of this type. However, generally, the degree to which innovation provides important benefits to the public external to the firm’s own income or expenses suggests a fundamental basis for identifying types of innovation that should be approached with great care under Alice. 111 To a degree, the Federal Circuit seems to have edged toward the principles offered here in its DDR opinion, creating an exception to Alice for software “necessarily rooted in computer technology in order to overcome a problem specifically arising in the realm of computer networks.” DDR Holdings, LLC v. Hotels.com, L.P., 773 F.3d 1245, 1257 (Fed. Cir. 2014). But this limitation is too general under the theory of this paper and might extend to online security systems that are functionally equivalent to general site security at a factory. In any event, the DDR decision ha