Lisa Larrimore Ouellette is a Professor of Law at Stanford Law School and a Senior Fellow at the Stanford Institute for Economic Policy Research. She is an acclaimed teacher and nationally recognized intellectual property law expert. She spoke with SSRN about incentivizing medical innovation, whether AI is a suitable legal teaching tool, and patent law.
Q: A lot of your work centers around intellectual property law and innovation policy. How did these become your main areas of focus?
A: My PhD training was in physics, but I realized that the research questions that most excite me aren’t just about science: they’re at the intersection of science and law and policy. The innovation ecosystem is complex: it involves researchers and firms operating within a web of legal institutions that shape knowledge production. But I’ve found that the legal frameworks that govern these institutions don’t always reflect how scientific research works in practice. Untangling these mismatches is the overall motivation for a lot of my work.
Q: How did you make that jump from physics to law? They seem pretty different.
A: They are different, but I think many of the ways I approach legal questions are with the mindset of a physicist: trying to make sense of complicated institutions, modeling incentives, and testing those models against real-world data. And many of the legal questions I’m interested in relate to how science operates in practice, including how scientists use the technical information in patents and how patents fit within the broader toolkit of laws and policies that affect scientists and other innovators.
Q: You are one of the authors (along with Jonathan Masur) on the free patent law casebook “Patent Law: Cases, Problems, and Materials,” and you recently produced the 4th edition. Talk a little bit about how this book came about in the beginning.
A: I was motivated by feeling increasingly uncomfortable with asking my students to pay over $200 for a casebook, and by the increasing divergence between my own sense of how best to present the material and the existing pedagogical options. I’ve been reading literature on the benefits of active learning and formative assessment since I was a physics grad student, in part because a number of prominent physicists have pivoted toward conducting rigorous pedagogical research. When I came to law school as a student, I was surprised at how rarely these methods were used. So, I recruited my co-author, Jonathan Masur, to work on this with me, and one of our key goals was to include a bunch of practice problems that lend themselves to active learning and formative assessments, including out-of-class assignments and in-class multiple-choice polls and small-group activities.
Another goal was bringing more conceptual clarity to legal doctrines that can be very confusing for students. That actually led to two follow-on law review articles that we wrote as we realized that there was some lack of clarity in the doctrine that we thought we could help explain.
Q: Are there any big changes that distinguish this new edition from the last?
A: Interestingly, patent law has been relatively stable in the five years since we wrote our casebook. From 1980 through the 2010s, there were substantial changes in patent law, including a lot of Supreme Court decisions and a major statutory revision, so teaching patent law now looks very different from a patent law course in 1980. But in the most recent years, the Supreme Court’s few patent cases have mostly had a much narrower focus on the kinds of topics that don’t get covered in a basic patent law course. So, I’d say the changes we made to our casebook are more things like improving the pedagogical quality by adding new practice problems or interesting lower-court decisions that apply the doctrine in a new way.
Q: You’re mentioning how some of those things have stabilized more in recent years: do you think there are any big gaps in the rulings that you’re still waiting to be covered?
A: Yes – this was part of the motivation behind the two doctrinal articles I mentioned. One’s called “Disclosure Puzzles in Patent Law,” and the other is called “Real-World Prior Art.” These revisit the fundamental doctrines underlying the patent bargain: an inventor gets a patent only if they teach the public how to make and use the invention, and only if the invention is truly new. If it was already being used or sold in the real world, they can’t get a patent. These principles of disclosure and novelty sound very simple, and they’ve existed since the first Patent Act of 1790. You would think that these basic doctrines would be very well settled at this point, but there are actually important conceptual holes that we’ve found, partly from good questions our students have asked us and from trying to explain the doctrine in the casebook. In our articles, we tee up these questions and highlight them for the courts, to say, “you should clarify this,” and we give the answers we think make the most policy sense.
Q: In one of the more recent papers you’re an author on, “Can AI Hold Office Hours?,” you and your co-authors analyzed the accuracy of answers to various law questions, as given by three different artificial intelligence platforms. Overall, the inconsistency and inaccuracies of the AI answers led to the conclusion that these tools would not be a suitable replacement for office hours. When putting together this paper and looking at those AI answers at a minute level, what – if anything – did you find surprising about the results?
A: There was a lot that surprised me. This project is different from a lot of my scholarship in that rather than being about how the law affects new technologies, it’s about how this key new technology affects legal education and practice. As I learn more about AI tools, I think that the legal academy needs to be thinking about the role of AI in our teaching and practice and doing it in an evidence-based way. Before this project, I wasn’t aware of efforts to test AI’s ability to answer questions about a specific legal text.
We basically gave three different AI models our patent casebook and then graded their ability to answer patent law questions. I had initially thought that these tools might be useful to law students as a study aid or a supplement to office hours. I was particularly optimistic about Google’s NotebookLM tool, since it was particularly designed for grounding answers in a specific text that you upload, and Google had advertised it as reducing the risk of hallucination because of that. Interestingly, NotebookLM had the worst performance of the three models we tested, and all three of them had a rate of unacceptable answers that was high enough that we don’t currently recommend them to our students.
Q: Having done different research on innovation, at what institutional level do you believe the most successful innovation happens?
A: Different institutions play different roles in the innovation ecosystem. A lot of my research has focused on the role of universities and funding for basic research, which has been critical to the development of all kinds of technologies, from new drugs and pharmaceuticals to modern computing and GPS. But the universities themselves are generally not the ones that are doing that final development and commercialization, bringing it to the public. Which is more important? I don’t know – each institution plays an important role. What’s interesting to me is thinking about the way they interact and the way different legal institutions structure incentives for all of these actors.
Q: What do you think were some of the biggest things that the COVID-19 pandemic revealed about innovation in the medical field?
A: There are a lot of things. One is that incentives matter. When we are willing to invest real financial resources, we can do extraordinary things. Before the pandemic, I had written about the low incentives for developing vaccines and why firms are less likely to invest in a vaccine than a corresponding therapeutic drug – in general, they’re much less profitable due to the legal institutions that structure incentives in that space.
But with COVID-19, the federal government poured substantial funding and other support into vaccine development, including by committing to purchasing successful vaccines at scale. That shift helped multiple vaccines reach authorization on a record-breaking timeline that experts hadn’t previously thought possible. I think that’s a major success.
It’s also a nice illustration of the way that the incentives for innovation and the way we allocate access to those innovations can be decoupled, which is something I had also written about before the pandemic. The vaccine manufacturers received large financial awards for their development, but at least every American who wanted a COVID-19 vaccine was able to receive one for free. The cost paid by patients and the price received by the manufacturers don’t have to be the same, and I think, in many cases, shouldn’t be the same.
I wrote about these and many other lessons from COVID-19 in a series of short essays with Rachel Sachs, Nicholson Price, and Jake Sherkow, which we consolidated into two parts on SSRN [Part I, Part II] and then synthesized into a book chapter, “Innovation Law and COVID-19.”
Q: Are there any ongoing projects or specific work that you’re doing right now that you’re excited about?
A: That’s a hard question because I always have a bunch of projects that are in stages of data gathering and brainstorming or drafting, with a number of different co-authors. I’m currently excited about two different projects that relate to patent disclosure. As we’ve already discussed, one of the fundamental requirements for getting a patent is that you disclose enough information for other researchers to make and use inventions, and I’m studying how well the patent system actually enforces this disclosure requirement. So far, I think the answer is “not very well.” It rarely gets raised by patent examiners who decide whether to grant a patent application. And when it is raised, it’s only a very narrow slice of disclosure problems, and not most of the problems that scientists would recognize as issues in patent documents. In both of these projects, I’m trying to quantify that a bit more in different contexts, and to think about ways to improve patent disclosures going forward.
Q: Is there any other work you’ve done that you’d like to highlight as particularly timely?
A: One of my favorite recent articles is “Valuing Medical Innovation,” co-authored with Daniel Hemel, which is about the many ways that incentives for medical innovation are misaligned from the value of those innovations. One problem is that we pay high prices for drugs with limited efficacy, and this issue of high drug prices has gotten a lot of attention. On the flip side, there is the more invisible problem of the innovations we’re missing out on that might exist under a more rational system.
I think this problem of missing innovations is particularly timely given the recent dismantling of federal support for biomedical research. Every rigorous estimate I’ve seen suggests that each taxpayer dollar we spend on innovation results in many more dollars of societal benefits – like economic growth and employment and health gains – and that the U.S. actually under-invests in science and innovation, compared to these benefits. But instead, we’re now cutting federal research funding, which will have massive costs to the U.S. and the world in terms of technologies that don’t get developed as quickly or at all. “Valuing Medical Innovation” is about a small piece of this problem in terms of incentives for pharmaceuticals, like how current Medicare and Medicaid pricing structures aren’t well aligned with social value. I think both that small piece and the broader issue of, “are we providing rewards that are commensurate with the value of the innovations we receive?” are both really important to think about right now.
Q: What kind of changes do you think would be a good start in encouraging the kind of medical innovations that provide the most societal benefit?
A: For one thing, policymakers should continue tackling the problem of paying too much for things that don’t have demonstrated value in terms of comparative efficacy, such as by demanding better evidence of efficacy before we’re using public dollars to, say, reimburse drugs through Medicare or Medicaid. But also, we should be willing to pay more for things that really do have a big impact.
I think the recent cuts to Medicaid go in the wrong direction here. Medicaid reimbursement is already lower than reimbursement for other covered populations, which creates a distortion such that there’s less incentive to work on problems that primarily affect the Medicaid population. Reducing Medicaid funding is just going to exacerbate that problem.
And we should not only think about how we pay for things on the back end, but also what we fund on the front end. For example, funding the NIH has been the driver of a huge amount of biomedical innovation in the U.S. and in the world. The recent cuts to NIH grants and the proposal for slashing government research budgets going forward are penny-wise, but pound-foolish: there may be some short-term cost savings for the federal budget, but the long-term cost to society will be enormous.
Q: What do you think SSRN contributes to the world of modern research and scholarship?
A: It’s very important to legal scholarship to have a mechanism for people to post and find papers before they’re published. There are endless debates over whether legal scholarship should involve more formal peer review, but I think it’s at least the case that most of the best law review articles do benefit from a lot of peer input before publication. SSRN is one of the mechanisms that facilitates that.
The main place I learn about new working papers is through SSRN ejournals, including papers by scholars whose work I might not otherwise have encountered. Sometimes I reach out to those scholars, or I get emails from other scholars who saw a new draft I posted on SSRN.
I also think it’s useful to have a place to post pre-publication papers because there’s often a delay between when a piece is done and when it is finally published. One example of that is I had a short piece, my “Trade and Tradeoffs” essay on international patent exhaustion with Daniel Hemel, where the Federal Circuit ended up citing the SSRN version before it was actually published.
More About Lisa Larrimore Ouellette
Lisa Larrimore Ouellette is the Deane F. Johnson Professor of Law at Stanford Law School and a Senior Fellow at the Stanford Institute for Economic Policy Research. Her research focuses on intellectual property law and innovation policy. She holds a J.D. from Yale Law School, a Ph.D. in physics from Cornell University, and a B.A. in physics from Swarthmore College. She has conducted scientific research at the Max Planck Institute, CERN, and NIST. Professor Ouellette is an acclaimed teacher and nationally recognized intellectual property law expert. She coauthored a free patent law casebook, “Patent Law: Cases, Problems, and Materials,” which has been adopted at over 70 law schools. In 2018, she received Stanford’s John Bingham Hurlbut Award for Excellence in Teaching. Her commentary has appeared in publications including the New York Times, Wall Street Journal, TIME Magazine, and Slate. She was also appointed to a committee of the National Academies of Sciences, Engineering, and Medicine to recommend strategies for better aligning medical innovation with disease burden and unmet needs. Prior to her appointment at Stanford Law School in 2014, Professor Ouellette was a Postdoctoral Fellow at the Information Society Project at Yale Law School. She clerked for Judge Timothy B. Dyk of the U.S. Court of Appeals for the Federal Circuit and Judge John M. Walker, Jr. of the U.S. Court of Appeals for the Second Circuit.
You can see more work by Lisa Larrimore Ouellette on her SSRN Author page here