We can distinguish three forms of “mixing” of funding.
(1) Two or more projects, each funded on different terms, in which their participants, having the freedom of the university, talk with one another, learn things, and apply what they learn in their own project. Thus, one project’s ideas (and findings) might “contaminate” another project, and so provide that project with “support” and in doing so might then be said to form a right by the sponsor of the first project to the patentable inventions made in the second project.
(2) A single project, funded by and with commitments to a single sponsor, that draws on resources (money, equipment, personnel) that have conflicting conditions on their use. For instance, a software development project that has agreed to be open source draws on a proprietary library of functions; or, a software development project that has agreed to be proprietary incorporates an open source library of functions that requires the entire code base to be open. Or imagine a project that has made a commitment to its sponsor that the sponsor will have the first right to negotiate a license to any patentable invention, but that project then uses a commercial instrument provided to the university on the condition that the company receives a royalty-free license to use any invention made using that instrument.
(3) A single project, funded by and with commitments to a single sponsor, in which investigators without the sponsor’s permission, farm out work to another organization that, to do the work, requires commitments that violate the commitments made to the sponsor. In this case, the investigators, rather than doing the work themselves, choose to involve others to do some of the work instead, without flowing down their sponsor’s requirements.
Each of these instances might be called “mixing” of funding sources.
In the first, the “mixing” is a matter of the open environment of the university, a university without firewalls. Let’s call it “open” mixing. Where a university accepts funding that forbids open communication, it is managed as “classified” research. While there are big policies about when and how to accept such funding, many universities agree to do classified research for the federal government, and for that purpose create separate, controlled laboratory space and require those working in that space to sign non-disclosure agreements. Those spaces, then, generally are subject to export law requirements involving what non-US nationals can receive information or work products produced under classified conditions. It’s not an easy thing to do, but universities do do it, and it’s a separate policy altogether. While the patent policy might be “uniform”–that uniformity has nothing to do with the management issue for compliance. Whether the university owns all inventions is simply not relevant.
A similar argument might be made for industry sponsored research agreements–if they carry requirements that would prevent publication or would create liability if the project drew on resources or shared ideas with projects that intended to be open, then those industry projects could be isolated as if they were classified. The non-disclosure requirements, then, might also trigger export control issues and only properly allowed or licensed students and employees might be permitted to work on such projects. At one point at least, the MIT Media Lab had such a policy–if a company wanted proprietary results, then its work was isolated from the rest of the lab and its resources and the company paid for every resource required to do the work. No mixing, no nothing, just access to specific talent for a specific task and then done.
Thus, for open mixing, common university ownership of patentable inventions is not relevant to the problem–the problem has to do with isolating research that’s not compatible with open mixing–or refusing to accept research on those conditions, even if that research support comes from federal sources. The problem is not who owns, but rather who delivers. Delegating the responsibility for delivery (or, rather, acknowledging that the university in managing the money never has a responsibility for delivery) is an effective way to manage open mixing. Individuals, then, have to know the conditions of support under which they work, and what they can and cannot discuss, or mix. Being in an isolated lab having signed a non-disclosure agreement and sat through a stimulating presentation on the forms of legal badness that will arise if one forgets one’s commitments for a moment is often adequate to the purpose.
The second form of mixing we might call “mine field” mixing. In such mixing, the university accepts (or allows the acceptance) of resources that are left out in the open for all use, but which carry obligations to the provider of the resource–money, equipment, lab space, even people who have been “seconded” or provided for a purpose. In one case, I dealt with a graduate student who was also an employee of a big tech company, which was paying his tuition. As an employee of the tech company, the grad student owed his employer assignment of any invention he made. This commitment made it impossible for him to join any research activity in which anyone else didn’t accept that a big tech company would be, necessarily, the joint owner of anything they invented together. The grad student was a little “mine” waiting for others to unsuspectingly trip over. Anyone might have similar obligations–a faculty member signing an overbroad consulting agreement, or a visitor-collaborator-volunteer arriving with prior commitments to an employer.
Another typical environment for mine-field mixing has to do with the receipt of resources for university use. Companies may place prototypes of equipment for evaluation or testing–with conditions on the use. Often the conditions include reporting of anything new done with the equipment and a license to any inventions made. Sometimes it appears that the requirement for reporting and a license is just there to prevent those providing testing to jump out ahead of the company itself and claim improvements or applications that then block up the market for the equipment. Bad behavior, but expectable even at universities, and so a requirement that gives the provider some degree of freedom to continue to practice.
But the more insidious mines are planted in license-in agreements for things like software and publications. Software is often purchased for “educational use only.” What happens, then, when such software is used in a project funded by industry, with deliverables directly resulting from use of the software to the sponsor? A mine. University ownership of a patentable deliverable does not resolve the issue. The problem is that the university has no authority to use the software to prepare a licensable deliverable for a for-profit company. The use is not “educational” but rather use “at an institution that also provides education.” The sponsor then gets a deliverable that includes a mine. Worse and worse if the deliverable incorporates outputs from the software or text from publications licensed for educational use only. Now think what happens if a university claims to own these deliverables and then of all things requires the sponsor to take a license to prepare a commercial product based on the deliverables. The ownership position might give the university the authority to grant the license (and to make commercialization demands), but ownership only increases the university’s liability with regard to the deliverable.
Patent ownership has nothing to do with avoiding liability for mine field mixing. It may well increase that liability. If one wants a “uniform” policy here, it must be directed at how people and products with obligations that conflict with an open environment interact with that open environment. One might even imagine a policy that forbids such obligations, unless those people and objects and services that carry the obligations are clearly announced and marked and separated in some way from the open environment, so there is no “mixing” though there may be salutations and other pleasant exchanges. Otherwise, every sponsor must be on notice that the only deliverables one might get from university-hosted research is what any individual can legally deliver, with whatever limitations that might arise based on who and what that individual has “mixed” with.
A prudent sponsor, then, will demand that deliverables are original, and that the preparation of the deliverables use only resources that are acquired for the project, or otherwise come with conditions consistent with the commitment made for project deliverables. A prudent university will then construct its research practices so that it manages the money but delegates the deliverables to those doing the work. Prudent investigators will accept only those conditions that they know they can comply with, knowing too that they personally are liable for failing to comply as they have agreed. This is exactly what the standard patent rights clause requires for federal funding, and exactly what university administrators have chosen, almost to a person, not to do. Displacing the responsibility for research deliverables to university administrators by means of a claim to own all such deliverables, then, creates double the liability–the investigators no longer have a personal interest and so don’t so much care what they do; the university administrators, by insisting that everything owned should be “commercialized” create even more liability for compromised deliverables. Pigpen administration, bringing its dirtiness with it, and then demanding even more money for the specialized skills to somehow mitigate all that dirt they’ve created.
The third form of mixing, however, is intentional. In intentional mixing, one takes support from one sponsor with one set of commitments, and then deliberately moves work to another organization with a conflicting set of commitments. A classic example is Heidelberger’s development of 5-FU, an anti-cancer class of compounds, at the University of Wisconsin. Heidelberger receives federal funding to support his work, but when he has characterized the prototypal compound, he sends the work to a colleague at a drug company to be synthesized there. WARF obtains the patent rights and signs an exclusive deal with the drug company, with no rights to the government. The Public Health Service throws a fit, sues WARF, and Wisconsin settles by assigning a share of the ownership interest in the patents to the federal government. The Public Health Service promptly licenses its share non-exclusively and royalty free. 5-FU becomes the gold standard for certain cancer therapies, but the claim is that its development was “delayed” by the PHS policy of making discoveries broadly available rather than choosing a single company to be responsible for the efficacious development of the discovery for the greatest public benefit (or, put another way, to monopolize the opportunity in whatever way seems most profitable).
Or consider a new faculty member hired from industry, where she has patents assigned to companies on technology that she wants to use in her university research. Beyond the opportunity for infringement, the problem of mixing has to do with the deliverable. Here, the deliverable comes with an intentional bomb–deliberately designing the deliverable so that it is unavailable for use by the sponsor. The sponsor may inspect the deliverable, but not make or use or sell it. What difference does it make that a university owns the deliverable as a required step in its delivery to the sponsor? What difference does it make that the university insists on “commercialization” of the deliverable? It would appear that both university ownership and commercialization create greater liability for the university and have nothing at all to do with compliance. If the university goes further and attempts to disclaim any “background” rights without revealing them, it creates even more suspicion that the university has not been diligent or may be seeking to profit from tainted assets.