Issues in Focus: Science and the Law
Reconciling Research and the Patent System
Working together, researchers and the Patent and Trademark Office can facilitate the development and application of knowledge.
Is tension growing between the goal of protecting intellectual property and the goal of advancing scientific and technological research? Some people think so. There's a perception in some quarters that the gears of the intellectual property system and the research establishment are grinding against one another; that the intellectual property system in this country may not be doing as much as it could to increase social utility with minimal transaction costs.
The concerns are many and varied. It is said that biological research will be disrupted by genomic patents; that academe will be damaged by the extension of copyright into the digital networked environment; that development of the Internet--and its promise for e-commerce, science, and civil society--will be retarded by business method patents; that proposals to protect investment-laden but uncopyrightable databases will hurt scientific research and commercial innovation.
In each of these examples, the worry is that the intellectual property system locks up new knowledge, in contrast to the goal of science, which is to gain new knowledge and disseminate it at little or no cost. These critiques are oversimplifications. The patent system is trying to ensure that the expansion of subject matter and the increasing number of patents do not impose unreasonable costs on society, research and innovation included. The R&D communities can help the system increase social utility for both the United States and the emerging economies.
Research and commerce
There are several reasons why people have recently perceived a collision between the goals of the intellectual property system and the goals of science and research. One is the fact that government-funded research is becoming an increasingly small proportion of our country's research budget. The Clinton-Gore administration has done a good job of fighting for increased funding for agencies such as the National Science Foundation (NSF) and the National Institutes of Health (NIH). In fact, the president's budget for FY 2001 includes nearly $3 billion more for science and technology than the government spent in FY 2000. Still, the trend of the past few decades has been that government funds less of the total U.S. R&D budget and the private sector funds more. In the 1960s, the government and private sector split R&D costs roughly equally. Today, two-thirds of those costs are underwritten privately.
The private sector looks for financial returns on its research investment, whereas Uncle Sam does not. Of course, the financial returns are very often packaged as, or linked to, intellectual property rights. Thus the increasing role of private funding in R&D has meant an increasing role for the intellectual property system.
In the 1970s, the government discovered that inventions that resulted from public funding were not reaching the marketplace because no one would make the additional investment to turn basic research into marketable products. That finding resulted in the Bayh-Dole Act, passed in 1980. It allowed universities, small companies, and nonprofit organizations to commercialize the results of federally funded research. The results of Bayh-Dole have been significant. Before 1981, fewer than 250 patents were issued to universities each year. A decade later, universities were averaging approximately 1,600 patents a year. Bernadine Healy, the former NIH director, has credited Bayh-Dole with development of the entire biotechnology sector. More than $30 billion of economic activity a year--250,000 jobs and more than 2,200 new companies--can be attributed to the commercialization of new technologies from academic institutions. Indeed, Bayh-Dole has worked so well that Japan is now seeking to emulate it.
In addition to the decline in the ratio of public to private research, a macroeconomic trend has brought research communities and intellectual property rights increasingly into each other's gravitational field. That trend is the growing economic value of knowledge. Competitive success in a market economy depends more and more on the knowledge a company holds, from the skills of its employees to the results of its latest research. It is estimated that the largest fraction of our gross domestic product, some 6 percent, is attributable to knowledge-related goods and services. It is also about 6 percent of all exports, which makes it the largest export sector as well, just ahead of agriculture.
As companies have come to realize that increasing value rests on knowledge, they have, naturally enough, pushed to convert that value into assets. One way--perhaps the principal way--that conversion occurs is through intellectual property rights.
This does not mean that the intellectual property system is intruding on research and science. It does not mean that the system for protecting intellectual property has locking up knowledge as its main goal, in opposition to the goal of the arts and sciences, which is the dissemination of knowledge. To the contrary, in the patent system, the basic quid pro quo is clear: dissemination in exchange for exclusive rights.
That is a point I want to emphasize. A properly calibrated intellectual property system balances within itself these two fundamental principles: protection and dissemination of knowledge. Information is a public good. If the information already exists, maximum utility is best achieved by distributing it at marginal cost. But the trick is the phrase "if the information already exists." The distribution principle is balanced by an information-generating principle: If you do not give people an incentive to produce information, there will be none to distribute.
John Barton has said that the "right price" for transferring information is the marginal cost of such transfer "unless there is a solid economic basis for an exception." In sectors where the initial investment can be high and the costs of copying are low, such as biotechnology, a solid economic basis does exist: the need to provide "incentives for research, authorship, and the like."
Practically every major policy dispute about intellectual property centers around the question of where to achieve balance between these two principles: the best way to generate intellectual property and the best way to distribute it. For that reason, our constitutionally based patent and copyright systems are founded on, and clearly provide, incentives for information disclosure and dissemination.
It is far, far better for a researcher to be working with funding from a company that seeks patented technology than with a company that would try to protect research results as trade secrets. The patent system not only permits the scientist to publish results, it also ensures that those results, in the form of the specification of a patent, will be published for all the world to see regardless of whether the researcher ever gets a single article past the reviewers. This disclosure, in turn, facilitates improvements in technology, allowing it to be expanded on while also building an important database of technological information. This disclosure incentive is the social contract on which the patent system is premised. In fact, the Patent and Trademark Office (PTO) has put a great deal of energy into promoting real, widespread disclosure of technology and technological ideas.
The best example is our completion last year of a two-year project to put our databases on the Internet. Now there are complete searchable texts of all registered and pending trademarks and of all U.S. patents granted since 1976, along with full-page patent images (drawings and schematics) to complement the patent text database. We have put on the Internet a two-terabyte database system of some 21 million documents. With budgeting luck, we can put the rest of our database up next year.
This can have an enormous beneficial impact on U.S. R&D. One of our international sister agencies, the European Patent Office, estimates that over $22 billion dollars a year is wasted on research that has been done before. Although repeating experiments might make for good high-school science projects, it's good neither for business nor for achieving tenure. Putting our patent databases online will help researchers understand more quickly what avenues they should pursue to achieve real innovation.
Another change in our patent system that advances the cause of disclosure is a provision in the new patent reform bill. It became law late in 1999, and I was pleased that the Clinton-Gore administration supported it. The new law provides for the publication of patent applications 18 months after the U.S. filing or priority date, unless one opts out by not filing overseas. This "pre-grant publication" will allow U.S. inventors to see an English-language translation of the technology that their foreign counterparts are seeking to protect in the United States and elsewhere at a much earlier point than today. As a result, it will allow people to better understand the state of the art, so they can improve on it and make wise R&D investment decisions.
New areas for patents
Despite this balancing that occurs between information dissemination and exclusive rights, there is no question that, once information exists, intellectual property rights impose costs on people who would use the information. For that reason, concerns have been raised about the expansion of patent subject matter; that is, applying the law of patents to emerging technologies or new areas of the economy.
One of the metaprinciples of our 210-year-old patent system is that it is technology-neutral. It aims to apply the same norms to all inventions in different sectors and technologies. Some people are critical of this uniform system. But the uniformity of the patenting standards of novelty, nonobviousness, and utility have allowed the patent system to respond to whole new sciences and entire new industries without the need for Congress to constantly retool the law.
This is not to say that each new invention in every new field of technology must be patented; that is the choice of the inventor or the owner. The inventor of the World Wide Web, for instance, chose not to avail himself of intellectual property protection. Moreover, it does not mean that every patent or copyright will produce license fee obligations for science and research. Many inventors and writers choose to make their protected works available to nonprofits for low or no royalties. But the general rules of the patent system mean that it evolves based on demand for protection, when researchers and those funding research believe that they need protection in order to secure a return on their investment of time, energy, and money.
The PTO plays a role in the evolution of the patent system. As an administrative agency charged with the application of the law as it exists, we take our guidance about what is patentable and what standards we use for granting those patents from Congress and the courts. We are receptive to a continued broad view of subject matter eligibility, where appropriate, because 20 years ago the Supreme Court instructed us that "Congress intended statutory subject matter to 'include anything under the sun that is made by the hand of man.'"
Although we are an administrative agency, we recognize that our examiners also serve in a quasi-judicial role because they are responsible for judging the patentability of applications that come before them. Similarly, the administrative patent judges on our Board of Patent Appeals and Interferences are responsible for handling appeals from the final determinations of examiners and for determining priority of invention.
Because neither the Supreme Court nor the federal courts issue advisory opinions, our examiners sometimes confront new issues and new problems with only the general guidance of existing case law to help them. To help administer this process more uniformly, we have recently been promulgating examination guidelines, including written description and utility requirements for subjects ranging from software to genomic inventions.
In the hundreds of thousands of patent applications that the PTO handles annually (270,000 in 1999) we inevitably confront new issues in emerging technologies. Whether or not we issue a patent provides an answer to the new question. However, the final resolution of the matter, which is sometimes a significant policy issue, may not occur until after adjudication through the federal courts, or in the rare situation that the Supreme Court takes a patent case, or the even rarer case that Congress decides to intervene. With the single exception of thermonuclear devices, however, Congress has never removed particular subject matter from the scope of the statute.
Sometimes the new issues are purposefully and carefully framed by the applicant, the examiner, and/or the board to provide appropriate test cases. This is one way in which the PTO promotes the evolution and maturation of intellectual property policy. By prompting legal decisions, it facilitates the application and interpretation of the federal patent statute. The recent utility guidelines relating to the examination of genomic inventions are a good example of this. This process of legal development is cumulative. It is evolutionary, not revolutionary.
Patents on software and business methods
Software and business method patents are an excellent example of the evolutionary nature of intellectual property policy. Responding to concerns about which aspects of software-related inventions should be eligible for patent protection, the PTO issued guidelines, first in 1989 and again in 1996, detailing proper analysis for computer-related inventions. The first set of guidelines recognized that although algorithms per se are not patentable, practical applications of mathematical algorithms may be. Building on the earlier version, the 1996 guidelines provided a uniform methodology for examining computer-related inventions and included the recognition that business method processes implemented through a software-based system may be patentable subject matter if they have a useful, concrete, and tangible application.
In 1999, in State Street Bank & Trust Co. v. Signature Financial Group, Inc., the federal courts validated the PTO analysis in those guidelines. The court also rejected the so-called "business method exception," stating that inventions of this nature may be invalid on other grounds, such as lack of novelty or obviousness, but not because they were improper subject matter.
This has been an evolutionary development. In keeping with the statute, we've been granting software patents for a quarter-century (about half of the time that programmable digital computers have existed) and what people call business method patents for a good 15 years. We've issued patents on methods of teaching since the 1860s. Moreover, the PTO is not the only patent office providing increased protection for computer-related inventions. Both the European Patent Office and the Japanese Patent Office have recognized that innovative aspects of software-related inventions may be patentable.
To those who are concerned about the ultimate ramifications of the State Street decision and who criticize our ability to examine business method applications, I say that we are working tirelessly to ensure that the PTO has the skills and resources we need to handle the growth in business method filings, which doubled between FY 1998 and FY 1999 from 1,300 to 2,600. To handle such a dramatic increase in workload, we obviously need more staff. So in the past two years we have hired more than 500 new examiners in the Technology Center that examines software, computers, and business method applications. Those examiners have an average of four years of practical experience in industry, and about 50 percent have advanced degrees. About 450 Ph.D. scientists examine our technologies. In addition to requiring examiners to have a scientific or engineering background, we also are recruiting candidates with business backgrounds.
This raises another issue: the quality of the work we do. We hear from some sharp critics, usually armed only with anecdotal evidence. I respectfully note that the PTO conducts the only regular comprehensive study of patent and trademark quality. We have reviewed our quality for years through the Patent and Trademark Quality Review process against six major quality measures, such as whether the best prior art was applied. These reviews are conducted independently of the Patent Office by seasoned Grade 15 patent review examiners whose sole job is this review function. This process reveals a consistent level of quality with no apparent deterioration over time. Even though this data is freely available, critics rarely consider it. This is not a case of the fox guarding the chicken coop. The Commerce Department's inspector general regularly reviews this process, and we certainly welcome any additional third-party scrutiny.
We also measure quality with the help of that most demanding group of third parties: our customers. The PTO's annual survey of customer satisfaction has shown fairly dramatic increases in customer perception of search quality, rising almost 50 percent in the past three years alone. These results are remarkable, especially given the steady drumbeat of anecdotally based criticisms to the contrary. This is not to say that search and examination are always perfect and cannot be improved. Quality management, after all, is a continuing process. But I am very pleased with the improvements we've made.
Our examiners have access to more prior art than at any other time in our history. Our in-house patent database and our commercial database provider give them access to more than 900 databases, including Westlaw, Lexis-Nexis, and Chemical Abstracts. From their desktop computers, patent examiners can also search the full text of more than 2.1 million U.S. patents issued since 1971, images of all U.S. patent documents issued since 1790, English-language translations of 3.5 million Japanese patent abstracts, English-language translations of 2.2 million European patent abstracts, IBM technical bulletins (a key software database), and more than 5,200 nonpatent journals. And our paper search files and libraries are still in place as well.
Finally, common wisdom to the contrary, examiners have had a consistent level of search and examining time; time that is adjusted for the complexity of the technology. Does that mean we should not devote even more time, if possible? Of course not. We should definitely try to devote more time to search and examination, especially for those technologies that are emerging or that are becoming more complex. Like most things, however, it is a question of resources.
The appropriations process over the past few "capped" budgets has resulted in the PTO being denied access to more than $200 million of its patent and trademark fees this year--about 20 percent of the total. This is a significant problem, and we hope to propose a permanent fix for this inequity to Congress.
It should also be noted that our original search and examination are not the end of the quality story. There are additional safeguards. Rule 56, for example, requires that each applicant disclose the material prior art of which the applicant is aware, under penalty of possible forfeiture of the patent. We are currently discussing additional means to heighten this obligation. There is also Section 301/Rule 501, which permits any art, even art submitted anonymously, to be placed in the file and used in subsequent litigation. Finally, there is the reexamination system, which now provides for reexamination in view of unconsidered or newly discovered prior art. Surprisingly few people avail themselves of either of these options.
The PTO commissioner may also order reexamination, as I recently did with a Y2K-fix patent, when the prior art and broad public concern warrant it. However, when expanding reexamination was recently debated in Congress, certain interests (principally independent inventors and the university research community) opposed it, apparently worried about its potential for abuse or hoping to maintain the fear of expensive and debilitating litigation as a barrier to entry. This resulted in a significant curtailing of the scope of reexamination.
Of course, as new technologies enter the realm of patentable subject matter, we need examiners with not only new skills and training but also with access to additional sources of technological information. That's why in 1999 we held hearings in San Francisco and Washington to solicit ideas on how to expand access to nonpatent literature. A number of organizations with access to such information, such as the Securities Industry Association, meet with our examiners on the state of the art and the databases that contain this information.
Roles for business and science
There are at least two ways in which the research community, both public and private, could help us improve the examination of software-related patent applications. First, there is a real problem in the software industry with commonly accepted terminology. One of the basic tenets of the U.S. patent system is that each applicant is allowed to describe his invention in his own words. In case law parlance, "the applicant may be his or her own lexicographer." The PTO then relies on the applicant's disclosure to determine the meaning of terms used in the claims. At the same time, our examination guidelines explain why it is important, particularly for prior art purposes, for applicants to use commonly accepted terminology. The software industry and the computer research community as a whole--ironically, an industry built on languages and focused on standards--should explore whether and how more common language in patent applications could improve the patenting process.
In addition to this nomenclature issue, there is also a place for industry-wide work in developing a robust, well-organized library of prior art. The 900 databases we now make available to patent examiners are an enormous improvement over the situation of 10 years ago, but those databases are still incomplete. A complete universal library of software prior art should be built, but the question is who should do it and how.
Some believe that the PTO should embark on this effort, but that is a daunting proposition. With other technologies, we rely heavily on private or nonprofit databases, such as Chemical Abstracts from the American Chemical Society and Medline from the National Institutes of Health. I ask the National Research Council's Board on Science Technology and Economic Policy to consider one definitive research project: to make recommendations on how we could move toward a universal library database for software-related inventions. We need help in exploring what resources we could bring to bear from organizations such as the Institute of Electrical and Electronic Engineers, NSF, and the private sector, as well as how we could divide up participation and the burden in such an effort.
Of course, the computer industry is not alone in its struggle with legal issues. Advances in biotechnology have sparked vigorous and emotional debate regarding the patenting of certain types of biotech inventions. At the heart of the controversy is the issue of patenting inventions concerning life forms and gene fragments that can be critical as research tools.
The patentability standard for biotech inventions that has guided the PTO since 1980 is that a product of nature that has been transformed by humans can be patented if it is new, useful, and nonobvious. Products made from raw materials that give those materials new forms, qualities, properties, or combinations are patentable, provided that they are supported by either a well-established utility or by an asserted utility that is specific, substantial, and credible; for example as a marker for a particular disease or for gene therapy. Under current law, genomic inventions are patentable.
Some fear that patents on gene fragments, such as expressed sequence tags and single-nucleotide polymorphisms, might retard basic research and that these claims will form an intricate licensing web that will impede their use in developing cures for diseases. The PTO is cognizant of these concerns, and we continue to take steps to ensure that patent applications in these areas are meticulously scrutinized for an adequate written description, sufficiency of the disclosure, and enabled utilities, in accordance with the standards set forth by our reviewing courts. In fact, commenting on our revised utility examination guidelines, Harold Varmus, former director of NIH, who was previously critical of these guidelines, recently stated that he was "very pleased with the way [the PTO] has come closer to [NIH's] position about the need to define specific utility."
In discussing genomic patents and access to research tools, it's important to distinguish between patentability and access. Although the need for a possible research tool exemption, presumably for basic scientists, is a valid topic for debate, it should not drive a narrowing of subject matter in order to create such a de facto patenting exception. There are often more traditional methods of dealing with difficult access issues, such as the Justice Department's antitrust guidelines. And if the issue is truly a research tool exemption, then that is what we should discuss, being mindful, however, that the distinction between basic research and applied research grows fuzzier every year. Moreover, research tools themselves can be of great commercial value. We must not forget that one university's access issue may be another university's critical licensing revenue.
How can the scientific community help us improve and refine the system? Changes are taking place in our patent system as a result of the recent enactment of landmark patent reform legislation. There are also areas where the scientific community can help us improve the system even more. The patent reform bill signed into law by President Clinton in November 1999 includes several changes in patent law that are an important step forward. These include:
- A guarantee of a minimum 17-year patent term for diligent applicants so that they are not penalized for processing delays and for delays in the prosecution of applications pending more than three years.
- The publication of most patent applications 18 months after the U.S. filing or priority date, unless the applicant requests otherwise upon filing and states that the invention has not been the subject of an application filed abroad.
- The establishment of a limited defense against patent infringement to inventors who developed and used a method of doing business prior to that method being patented by another party.
Of all the bill's substantive provisions, the pre-grant publication of patent applications represents the most significant step toward information disclosure and global harmonization. Still, much more remains to be done.
A truly global patent system will be realized only when the currently divergent substantive and procedural requirements for the granting and enforcement of patents converge. The costs associated with maintaining our current duplicate systems cannot be tolerated much longer. A month-long meeting of the World Intellectual Property Organization (WIPO) in Geneva just culminated with the adoption on June 2 of the Patent Law Treaty (PLT). The treaty harmonizes patent procedures around the world in order to reduce the high costs of securing patents in multiple countries. The PLT will come into force once it has been ratified by 10 WIPO member states.
At the same time, much needs to be done about differing substantive requirements. For example, the United States alone grants patents on a first-to-invent basis, whereas the rest of the world grants patents on a first-to-file basis. We also have an important and generous one-year grace period that benefits our patent applicants. Earlier attempts to harmonize these and other differences were fraught with controversy and met with failure. Now that there has been a cooling-off period, I ask for the support of the scientific community in studying the impact of the changes here, or throughout the world, that would be required to achieve such a global patent system. We need the benefit of your perspective and the rigorous academic analysis of the issues that you could offer.
One of the most obvious areas where the research and scientific community can help us is in issues concerning the transfer of knowledge and technology. We need the U.S. scientific and research community to be involved, including helping researchers and scientists in other countries understand the impact of these proposals. This partnership could greatly aid developing countries in particular as they move forward into global economic development.
The impact of intellectual property rights on technological innovation in our economy is clearly of the utmost national and international significance. These issues deserve the attention of us all.
Q. Todd Dickinson is assistant secretary of the U.S. Department of Commerce and commissioner of patents and trademarks.