I asked ChatGPT for “the key points of the Bayh-Dole Act.” Here’s the first response:
The Bayh-Dole Act is a United States federal law that was passed in 1980. It allows businesses, universities, and other organizations to retain ownership of patents on inventions that are developed using federal funding. This means that these organizations can choose to license the inventions to other companies or to produce the inventions themselves, rather than having the patents automatically assigned to the federal government. The act was intended to encourage innovation and to stimulate economic growth by providing a financial incentive for organizations to invest in research and development. Some of the key points of the Bayh-Dole Act include:
- It allows organizations to retain ownership of patents on inventions developed using federal funding.
- It provides a financial incentive for organizations to invest in research and development.
- It encourages innovation by allowing organizations to license their inventions or produce them themselves.
- It stimulates economic growth by making it easier for new technologies and products to reach the marketplace.
- It allows the federal government to use patented inventions for government purposes, such as national defense, without requiring the organization that owns the patent to grant permission.
Let’s work through this.
The Bayh-Dole Act is a United States federal law that was passed in 1980.
Yes!
It allows businesses, universities, and other organizations to retain ownership of patents on inventions that are developed using federal funding.
Notly. Bayh-Dole is restricted to small businesses and nonprofits. Large businesses are addressed by executive branch patent policy–namely Reagan’s 1983 Memorandum. Bayh-Dole requires federal agencies to allow small business and nonprofit contractors to retain title to inventions (not just patents) made in work that has received federal funding. Title to inventions means that patents will issue to the invention owner. But owning inventions is broader than any patents that issue, since whoever drafts the patent application is at liberty to leave things out of the specification and claims that are still inventive, and patentable.
The difference between inventions “developed” with federal funding and inventions made in work that has (also) received federal funding is significant. Bayh-Dole’s implementing regulations make the point that separate accounting does not determine whether an invention comes within Bayh-Dole’s control. One could otherwise easily switch something that might become patentable to a non-federal budget and claim it was not “developed” with federal funding. That’s what USC tried to do in Mine Safety Appliances, but the court rejected the scheme. It is not an invention that is developed with federal money but that in research or development work for which a contractor has requested federal financial help, even in part (or, early or later in that work) an invention has been made and acquired by the contractor. “Developed using” and “made in work also funded by” are very different scopes of interest.
This means that these organizations can choose to license the inventions to other companies or to produce the inventions themselves, rather than having the patents automatically assigned to the federal government.
Garble. Contractors that elect to retain title to inventions they have acquired must file patent applications, and if they don’t timely achieve practical application (use with benefits available to the public on reasonable terms), then the government can request that they grant licenses (on reasonable terms) or the government can do so itself (and so the contractor would lose any royalties in the deal). It’s not so much, then, that a contractor can “choose” to license an invention under its patent rights or make/use/sell product based on the invention–it has to, or the government has the right (under federal contract, as required by Bayh-Dole) to require licensing. Of course, the “produce the inventions themselves” is just garble, something that one might find on a badly drafted university licensing office page.
The last clause of this sentence is just wrong. There is no mechanism in Bayh-Dole for patents to “automatically” be assigned to the federal government. There is also no mechanism for inventions to be automatically assigned to the federal government. Before Bayh-Dole, the federal government followed the Nixon variation of the Kennedy patent policy, codified as the Federal Procurement Regulation. The FPR prescribed the contracting instruments to be used for federal research. One requirement of the FPR contracts with contractors not eligible up front to retain ownership of inventions made in federal work (such as nonprofits or contractors doing research directly concerning public health) was that the federal government would acquire ownership of inventions made under contract unless the federal agency funding the work decided otherwise. If the agency did not decide otherwise (following a “greater rights determination” procedure) either the contractor or the inventor would assign the invention upon request. No automatic assignment.
Given that if the federal government acquired title to an invention, its policy was (following Kennedy/Nixon, which followed the 1947 AG’s recommendations) to release the invention to the public via “dedication or licensing” (depending on whether the government issued a patent to itself). The contractor in such cases still had full access to the invention and could do whatever it wanted by way of use and development, and if development, then should there be any additional inventions made outside the scope of the federal work, any such inventions could be owned and patented by the contractor. In this way, federal open access encouraged contractors to get a head start with inventive improvements, applications, combinations–any of which might be patented and thus provide “incentives” for continued development, all without having to hold exclusive rights in the invention made in federal work.
In a very real sense, then, Bayh-Dole removed that “free competition” incentive for a contractor with early knowledge of an invention (because, hey, their people made the invention) to move fast to obtain patents on improvements and the like and replaced it with an incentive to take one’s time about development while preventing anyone else from practicing any aspect of the invention made in federal work. The incentive for anyone other than the contractor now is–because they will not get access to the invention–to avoid the federal invention, design around it, attempt to block with patents lines of development, or to exclude the invention from developing industry standards. Pretty grim outcomes, reflected in the massive nonuse of Bayh-Dole inventions claimed by universities (where are the outcomes for those 50,000+ US patents citing federal funding in the Bayh-Dole era, held by universities?)
The act was intended to encourage innovation and to stimulate economic growth by providing a financial incentive for organizations to invest in research and development.
Nothing in Bayh-Dole states an objective to “encourage innovation” or to “stimulate economic growth.” These are just made up–but then, the whole account is just made up! Whoa on that. There is a way to get to innovation from practical application of an invention, but it takes a few hops. Put it this way: an invention may not be an innovation, even if practically applied. Someone uses the invention, but that’s not necessarily innovation–in many of the definitions of innovation that have been floated. Not Rogers’s “something new to an adopting group.” Not OECD’s “a new or improved product process that differs significantly from the unit’s previous products or processes and that has been made available to potential users (product) or brought into use by the unit (process).”
The intent of Congress–the only intent that matters–is set forth at 35 USC 200. There, we find that the policy and objective starts with using the patent system to promote the utilization of inventions arising from federally funded research or development. It’s use of inventions, not directly or immediately innovation or economic growth. Use of an invention could easily wipe out a whole segment of industry in the US with manufacturing done elsewhere, and the stimulation of economic growth, at least in the US, would take a hit. The other objectives stated by Bayh-Dole include maximum participation by small businesses, collaboration between nonprofits and companies, promoting free competition and enterprise (a pitch for non-exclusive licensing and limited exclusive licensing), US manufacturing, rights to the government for its use and to protect the public from nonuse and unreasonable use, and to lower the costs of administration of inventions made in federally funded work.
It is not then, merely a matter of a financial incentive–and financially incentives are nowhere to be found in Bayh-Dole. Even the use of the patent system is constrained so that financial incentives to invest in development of inventions arising from federal work are limited to “reasonable terms”–not patent monopoly pricing, say, but rather the price one would expect if there were competition, even if that patent on an invention made in federal work suppresses for a time that competition. Thus, a whole lot more has to be laid out to explain this idea of a financial incentive to attract “investment” in development. Nothing in Bayh-Dole has anything to do with giving anyone an incentive to invest in research. Federally supported research generally is intended to fully cover both direct and indirect expenses incurred by contractors requesting federal support for their work. There’s usually a shortfall on the order of 20% for universities doing a lot of federal work, but it is a long stretch to claim that the idea of gaining ownership of inventions made in that work works as a primary incentive–or any incentive at all–to do research. Just nice sounding garble.
Key points:
1. It allows organizations to retain ownership of patents on inventions developed using federal funding.
As the US Supreme Court ruled in Stanford v Roche (2011), Bayh-Dole allows contractors to retain the ownership of inventions to which the contractors have acquired title by conventional means. “Retain” cannot be used to suggest that contractors have title automatically, as if title conditionally vests with contractors and not inventors until contractors decide they don’t want title or screw up their patenting work.
2. It provides a financial incentive for organizations to invest in research and development.
No financial incentive. Just a limited use of the patent system. A reasonable “incentive” to license patents for money (if one is a university) or to try to practice the invention and turn a profit on sales at some point (if a small company). Nothing about investing in research. Oh, and for a lot of inventions–such as methods and research tools or new varieties of apples–there is really not much “development” that has to be done.
3. It encourages innovation by allowing organizations to license their inventions or produce them themselves.
Innovation is also encouraged by open access–creation of platforms and libraries of tools, interoperability of parts and devices, standards formation, and the like. Using patents to suppress alternatives, competition, and collaborations does not necessarily encourage innovation at all. So this is just a random claim without context, not supported by the law’s own statement of policy and objectives.
4. It stimulates economic growth by making it easier for new technologies and products to reach the marketplace.
Bayh-Dole adds more bureaucratic red tape. Inventions acquired by contractors must be disclosed to the federal government in a written report, there are timelines for electing to retain title, and to file patent applications, and provide reports on utilization of inventions, along with restrictions on licensing and assignment and compulsory licensing if various requirements for utilization are not met. None of this matters for promoting use of inventions made in federal work.
5. It allows the federal government to use patented inventions for government purposes, such as national defense, without requiring the organization that owns the patent to grant permission.
No. Bayh-Dole requires federal agencies to use a patent rights clause under which contractors agree–as a condition of obtaining the federal funding they have requested–to grant to the federal agency a license to practice and have practiced the invention–essentially, for all government purposes (following the language of the Kennedy patent policy from which this provision in Bayh-Dole was lifted). Practice means “to make, to use, to sell”–so it is not merely that the license grants the government a right to “use” but rather that the government has most of the rights of ownership, but for the right to grant exclusive licenses and the right to enforce the patent. That’s a much broader license than ChatGPT’s summary. So, oddly, the contractor has granted permission, not that the government gets to use without a grant of permission. For government use of patented inventions without permission of the patent holder, see 28 USC 1498.
What did ChatGPT leave out of its key points?
That Bayh-Dole repudiates the 1947 AG’s recommendations and authorizes federal agencies to deal in exclusive licenses, take a money interest in licenses, and sue US citizens and companies for infringement, or task chosen companies to do so on behalf of the government. That’s pretty huge.
That Bayh-Dole limits nonprofit assignments of subject inventions, requiring federal approval for assignments other than to firms that primarily manage inventions, and in any case requiring the assignee to follow the nonprofit’s patent rights clause.
The nonprofit patent rights clause places special restrictions on nonprofits in addition to the limitation on assignments–including sharing royalties with inventors and using any income earned with respect to a subject invention, after recovering only expenses incurred in the administration of subject inventions, for scientific research or education. That’s nothing super difficult for universities (though they mostly don’t comply and recover more expenses than allowed), but when a university assigns an invention to a company (as almost all do, using a broad exclusive license of all substantial rights for the term of the patent), the company must behave as if a nonprofit–but none do, and federal agencies turn a blind eye. Still a key point that gets lost because it is inconvenient to the folks who spread all that text out there making it appear that Bayh-Dole gives them a free hand to seek monopoly profits on anything they can, but especially medical treatments. Cruel and unusual folks.
What might we learn from this use of ChatGPT? Perhaps that it repeats gaslighting text indiscriminately, assuming that what’s out there in spades must be correct in spades. But that’s a bad assumption for a lot of areas. What we get then is that ubiquitous narratives become aggrandized as fact and technically accurate ones, or rational ones, or contrary ones, or alternatives get demoted. Sort of like how Google search operates, but with the algorithmic basis for making grand things grander buried even deeper in the operation of the technology.