NIST convened a symposium in April 2018 titled “Unleashing American Innovation”:
On April 19, 2018, thought leaders from across government, industry and academia gathered in Washington, D.C., to discuss how we can work together to address systemic barriers to catalyze the full potential of American innovation. The United States invests about $150 billion each year on federal R&D, but are the American taxpayers reaping the full benefit of that investment?
The premise of the symposium, then, was that things are not working all that great–there apparently are “systemic barriers” and that somehow we have failed to “catalyze the full potential” of “innovation.” Somehow, American innovation still needs to be “unleashed.” We have questions, then, about what the leash is made of, how it has been attached to the collar of American innovation, who has done the attaching, and who now will do the unleashing.
If the NIST language used wasn’t so full of bombast, one might think there was a serious problem. As it is, we find a serious problem couched in the bureaucratic language of self-congratulation: “We are doing tremendous work, so proud of it all, and on the chance we can do even better, let’s assemble ourselves for a big group pat on the back.”
Here’s a second paragraph like unto the first, so you can see the rhetoric in action:
The Unleashing American Innovation Symposium was about highlighting the nation’s journey to a new level of innovation performance.
We are already at a level of innovation performance, and we seek a “new level” of that “performance” worthy of “highlighting.” No blame here. No one will be called out. Everyone gets a trophy. T-ball for technology.
This path will seek the best models and approaches for converting the results of federally funded R&D and intellectual property into new companies and jobs as well as entirely new industries that bring new products, technologies and better healthcare to the American people.
So much going on here. The stated goal here for federally funded “R&D” and “intellectual property” is new companies, jobs, and industries. And these new companies will then do the work of innovation. Consider: it might be clear how a company might “bring a new product . . . to the American people.” But how does one bring a “new . . . technology”? Wouldn’t making a new technology–different than a product–available to the American people mean something along the lines of “releasing the technology to the public domain” or “licensing non-exclusively and royalty-free”? But reading any sort of logic into bureaucratic bombast is always a dangerous thing.
Finally, think a bit about the search by bureaucrats for “best models and approaches.” What could that possibly mean? How would bureaucrats know what constitutes a “best model” for “converting” anything? Why would there be a “best model”? Why not many models? What if the best model is to have no best model and be purely opportunistic? And if there is a best model, is it best only relative to the circumstances of a given institution, the inventions involved, and the industries that exist or are yet to be created? That is, what if there is a “best model” for inventions in medicinal chemistry directed at the pharmaceutical industry and a competing best model for those same inventions if directed at an industry that has yet to be created? In such a competition, what would public policy favor if the goal of federal “R&D” is to serve both existing companies while offer the prospect of new companies in new industries? Which feels more “innovative”? Which has the greater need to be “unleashed”?
We might go further–is there a “best model” for software? Open source? Proprietary source? Free/GPL? Public domain? Perhaps only bureaucrats would insist that software (such a huge category of activity) would be best served by a “best approach” other than that people should use their judgment about each situation. Gosh, imagine a whole symposium in which bureaucrats tried to articulate the idea that laws, regulations, and policies in matters of innovation from research ought to favor the personal judgment of those undertaking the research. I know–it would never happen.
The symposium also explored barriers that limit industry’s access to federal R&D and ways to maximize the economic, security and societal benefits to the nation.
A patent is in its default a barrier to limit “industry’s access to federal R&D.” Period. A patent licensed exclusively necessarily excludes the rest of industry in favor of a single player. An exclusive license agreement that precludes that licensee from granting royalty-free, non-exclusive licenses, or cross-licensing, or dedicating essential claims to a standard is a license that precludes “industry access to federal R&D.” A single company gains access, not “industry.” No question–the prevailing model of university patent licensing practice creates barriers that limit industry’s access to federal R&D. It does not have to be this way, but it is. That’s the reality.
We might characterize university patent licensing practices as a monoculture of “best practice.” This “best practice” involves demanding ownership of all inventions made by personnel at a university, filing of patent applications, and seeking to license exclusively all patents that issue. (Yes, a university will file patent applications on inventions made under an industry grant or in an industry consortia, even though the sponsors get a royalty-free non-exclusive license–but the university won’t then offer all non-sponsors that same non-exclusive license.–That is, the university does not offer industry access–only sponsor access.)
And what happens in a monoculture where every nonprofit patents and seeks to license exclusively? Consider carbon nanotube research funded by the NSF. The NSF spreads its basic research funding around among various universities. As a result, since 1996, 33 different universities hold 108 patents reciting NSF funding on carbon nanotubes. At about 30 claims per patent, for NSF-funded carbon nanotube research alone we have over 3,000 distinct claims covering many aspects of carbon nanotubes. Anyone in industry wanting to use a carbon nanotube has to navigate those 3,000 claims–and may need access to hundreds of claims across the patents of multiple institutions. I once dealt with a company that stated with some pride that it had obtained licenses from 20 universities for its nanotechnology, and that it was unlikely that any other company could possibly follow in its tracks.
Now tack on that each of those universities intends as its “best model” to grant exclusive licenses. Either they all license to the same company, or they fragment any future potential product that involves patents from more than one university into a patent gridlock–each university licensing to its own favored company, but no company having sufficient rights to create a meaningful product, and each licensee prevented from cross-licensing or dedicating their licensed rights to a standard.
Consider NSF-funded carbon nanotubes again. Rice University, where Roy Smalley, one of the leaders of carbon nanotube research was located, licensed its patents to Smalley’s startup. A proud moment for AUTM’s count of startups created with research patents. But Smalley’s startup failed to commercialize anything and sold itself to a holding company of a patent troll that aimed to acquire a monopoly on nanotechnology inventions. That is, a company that has it figured out–that industry will need a bunch of inventions to do anything, and rather than universities making NSF “R&D” available to industry, it turns out that instead the universities make their inventions available to organizations set up to exploit industry access for profit. Call it speculation rather than investment. Call it adding another “middleman” to the stack of costs, even if that middleman argues it is performing a service to industry to make up for the awful defects of university licensing practice. Meanwhile, the patent troll, having held the Smalley patents for some time in its front company, sold the patents to a Korean company (technically, transferred the exclusive licenses that amounted to assignments of the invention), which apparently holds the patents as a speculative investment.
This is the prevailing “best model” of university patent licensing practice. There’s a trail of inventions, patents, licenses, startups, and money made by universities and patent speculators–but virtually no access by industry, no commercial products, no benefit to the American public, unless you really are going to argue that benefit to patent speculators is the intended benefit to the American public.
One can repeat this analysis for any number of areas of federally supported research. Stem cells, for instance. Consider WARF’s management of a basic stem cell technology, licensed exclusively to a startup company without clearly reserving rights for nonprofit use. A “symposium” was called by Stanford to deal with WARF’s practice of the “best model” for university patent licensing–the exclusive deal, reserving rights apparently only for the University of Wisconsin. From that symposium came the “Nine Points to Consider” document, which argues for non-exclusive licensing–and which many universities then “endorsed” while ignoring altogether non-exclusive licensing (even though the Cohen-Boyer patents for gene-splicing were a great success and were licensed non-exclusively for a very low royalty–and according to Niels Reimer the Cohen-Boyer inventions would have been used by industry without any patents whatsoever). WARF’s best practice then screwed up not only industry access to a new technology but also university research access.
Let’s add one more layer, so you can see how screwed up the university best model is. Even once WARF sorted out its exclusive license (it required litigation to do so), all the universities got was an agreement that WARF and its exclusive licensee would not sue them for infringement. But even then, anyone doing research using WARF-owned inventions and making improvements would have nowhere to go with those improvements if they were within the scope of the WARF-owned patents. WARF or its licensee/assignee company would have to approve the use of these improvements outside of the university in which they were developed. Again, a huge barrier to access for industry–and for medical professionals.
It is *not* that academics don’t have access, somehow, to the WARF-owned inventions for research–it’s that there’s nothing much that can come from that research without WARF’s exclusive licensee/assignee’s approval. Surveys of academic researchers–“Do you have trouble gaining access to patented research tools”–are asking the wrong question.
That is, the research either must be dedicated to finding a way around the WARF patents (that is, find ways *not to use* an invention) or the work is done for the benefit of WARF’s exclusive licensee/assignee. Federal funding is either wasted or done for the benefit of a company with a patent monopoly. How, in what crazed mind, is this innovation “unleashed”? Clearly, to “unleash” innovation, one would get rid of the patent monopoly on the inventions from basic research. Anything worth developing will involve additional inventive work. The foundational inventions–the critical mass of cumulative technology, the platform, the library of tools–has to be open, whether that’s public domain, commons, standard, non-exclusively licensed on fair, reasonable, and non-discriminatory terms.
Consider disease diagnostic assays. Better diagnostic assays might lead to clearer diagnoses, better treatments. With many of these assays, any good lab medicine department can implement the assay from published information. Methods and materials. Nothing needs to be “developed” as a commercial product for these assays to be used. Any commercial product is for less capable clinicians or clinicians that lack access to a good lab medicine department. A commercial version does not necessarily mean a better version–rather, it means a mass-produced product that anticipates less capable users and tries to find a sweet spot of functionality that addresses the needs of the market while generating a profit, of course.
In the case of diagnostic assays, the effect of a patent position is to prevent expert practice in favor of a lesser commercial product that even experts are forced to use. Of course, one can also use a patent position to extract a “royalty” from those using the claimed invention–even if their use is not the same as the use by the commercial product on offer. In that case, expert users must pay to use a version of the invention that the patent owner is not using and has not developed for use by others. That is, the patent owner has added no value whatsoever to such an invention–the patent monopoly did nothing to attract private risk capital to develop the form of the invention used by experts. Such a patent serves no public purpose whatsoever for all those claims that are never worked by the patent owner–the patent merely allows someone to troll expert uses without investing in supporting those uses.
The premise of NIST’s symposium, then, is one of trying to figure out how to use the patent system to “unleash” the very thing that the present nonprofit use of the patent system comprehensively and arbitrarily leashes. One would think that the primary method of unleashing needed here would be to get nonprofit administrators out of the patent ownership business or use the patent system to expand the public domain or technology commons. But no, that is not what we get. Instead, platitudes disconnected from practice.