“Over the last several years, artificial intelligence has been the focus of patent offices around the world…. Perhaps a similar reflection will be performed as brain interfacing technology becomes increasingly powerful and more frequently used.”
Currently, brain recording and/or brain stimulation is used almost entirely for medical or research purposes. Invasive surgery is generally required to read neural signals with high temporal and spatial resolution. High resolution of neural signals enables researchers to decode a brain’s underlying intentions, sensations, reactions, etc. rather reliably, if only in constrained environments. With regard to stimulation, researchers have demonstrated the ability to trigger different types of effects. For example, stimulating the reticular formation can cause a subject to become awake; stimulating the amygdala can evoke fear; stimulating the motor cortex can trigger movement; stimulating the visual cortex can cause a subject to see light flashes. However, the state of the art is neither able to decode complex and unrestrained thoughts nor cause a subject to have a particular “target” complex thought. At least not yet.
Despite these limitations, some researchers have successfully triggered stimulation of a subject’s brain from information read directly from multiple other subjects’ brains. For example, signals from multiple “sender” brains were collectively processed and then transformed to stimulate another “receiver” brain. Such networks of brain signals are aptly dubbed “brainets.”
Many entities are hypothesizing and/or supporting ideas that neurotechnology will continue to undergo drastic advancements and support nearly unimaginable use cases. For example:
- The Royal Society predicted that – by 2040 – it was probable that neurotechnology would support “hands-free control of computers” and speculated that people could “become telepathic to some degree … through access to each other’s thoughts at a conceptual level”;
- Similarly, an international group of authors affiliated with 18 institutions (including Lawrence Berkeley National Laboratory, Duke University, Purdue University, I.M. Sechenov First Moscow State Medical University and Monash University) published an article predicting that – within 20-30 years – neuralnanorobotics would be administered to humans and connect to a remote computing system, so as to “serve as a personalized conduit, allowing persons to obtain direct, instantaneous access to virtually any facet of cumulative human knowledge”; and
- The government is betting on it. DARPA funded six teams to attempt to secure noninvasive or minutely invasive interfaces having high temporal and spatial resolution (50-ms temporal resolution and 1-mm3 spatial resolution) to support closed-loop recordings and stimulation from 16 or more brain regions
What if these scientists’ Borg-like prediction is accurate and DARPA is successful? Perhaps it is not far-fetched. The above-mentioned medical and research brain-machine interface (BMI) equipment is miniaturizing year after year even without DARPA’s help. Advancements akin to reading thoughts may not be far behind. Last year, in one study, neural signals were recorded while subjects read aloud sentences. The neural signals were processed to generate synthetic speech. Sentences confined to a list of 25 words were perfectly transcribed 43% of the time. And it scaled well. Doubling the number of words – which for conventional computers increases the difficulty exponentially – only halved the accuracy. Advancing from interpreting sentences to interpreting and mixing thoughts may be on the horizon.
If a time arises when people can opt into a system where their neural signals become part of a collective whole, if our thoughts can meld together, then what will this mean for our sense of individuality, originality, and industriousness?
How will we protect the creations that we author and invent, the hallmarks of our creativity and ingenuity? What might this mean for our intellectual property systems? Would current patent requirements make sense? Would the existing system still serve its purpose of “promot[ing] the progress of science and useful arts”?
So, let us embark on a thought experiment to explore the extent to which current patent system tenets may remain sensible in this potential future.
How Might Brain Interfacing Affect Inventorship Determinations?
In U.S. patent law, inventorship is attributed to those who “conceive” of an invention. It is straightforward for one inventor working alone or multiple co-inventors working in-person. Yet in our modern world of telephones and videoconferences, emails and instant messaging, it can sometimes be difficult to pin down.
Fortunately, each of those channels of communication offers a few advantages for determining conception: a defined start and stop of each utterance; and a view of who said it. Over telephones and videoconferences, people typically take turns speaking, and so there is a start and stop of what each person says. Similarly, emails and instant messages are discrete assemblies of words that go back-and-forth, coming and going from each individual. By default, our electronic devices automatically tag who sent them. Therefore, the people who were there, or on the message thread, can generally determine who came up with what.
The advantages disappear if multiple people’s thoughts are melded as computer bits on a server to conceive an invention. The start and stop of what each person contributed may be lost. The view of who “thought” of which concepts or parts thereof may be even murkier. In these realms, it is not clear whether the word “conception” would even apply, as the word historically has meant that a thought, at some point, was distilled and identified within someone’s wet brain, not assembled together in a computer.
Compound the above with the possibility that many, many people may share their electronic thoughts on a brainet. As the number of people increases, so does the burden of identifying inventorship. Even if one could track and attribute each person’s thoughts to a conceived idea, at some point, this process becomes nearly intractable due to the volume of people potentially contributing.
Currently, any person who materially contributes to at least one claim in a patent application must be listed as an inventor (MPEP 2109). That fundamentally assumes that there are a fairly limited number of people collaborating to the point of conceiving of an invention. Even in large, multi-disciplinary research institutions, rarely are there more than 15 or so inventors on a patent application.
A thought-melding computer could conceivably handle a much scaled-up version of collaboration. For context, massively multiplayer online games host thousands of players at any one time. There may be few, if any, practical limits on the number of individuals who can share and consume thoughts at once. Thus, should conception be limited to those who contributed “substantially” to a claim? Should there be a maximum number of inventors in any one application? Should our current patent system need to evolve, much like tort law that spawned the concept of class actions, to compensate for industries that “substantially” touch, or take from, many?
Anonymized thoughts (originating from unidentified patrons) may, potentially, also be part of conception in a thought-melding computer system. By definition, anonymous contributions to an endeavor are difficult to attribute to their origins. It may be impossible to attribute conception to all worthy inventors from whom thoughts originated and were ultimately melded. Further, the number of anonymous thoughts may scale far above whatever reasonable limits one may have in the existing patent system. Should there be provisions for anonymous John or Jane Doe inventors? Should there be an opt-in or opt-out registration period to claim inventorship for such anonymous thought contributors?
How Might Brain Interfacing Affect POSITA Characterizations?
Currently, a person of ordinary skill in the art is deemed to be a “POSITA”: an idealized expert whose intelligence is the standard for several aspects of patent law. For example, the written-description requirement is based on whether an invention is sufficiently explained such that a POSITA can practice it without undue experimentation. A POSITA is also the perspective to be used when determining whether an invention is obvious, and the POSITA is presumed to have universal knowledge of relevant prior art.
If it becomes commonplace for many thoughts to be melded together, then should the definition of a POSITA change as well?
A melding of multiple persons’ thoughts availed by brain-machine interfaces may result in substantially greater expertise than anything that existed previously. A collection of thoughts from a crowdsourced group of individuals may encompass a tremendous amount of information, potentially much more so than encompassed by any one existing human. A computerized melding of thoughts may be closer to an idealistic POSITA than anything else even contemplated.
Yet the existing POSITA has its faults. It is easily led astray from a design by a teaching away, and it cannot recognize critical values within a previously disclosed range that were not identified previously. A melding of thoughts may be more robust than a present POSITA. Should the POSITA also adapt to think more like a group or a melding of thoughts from several people? Should the bar for what is nonobvious be raised such that it impresses even a group of individuals melding their thoughts? Should those very thoughts, undoubtedly stored on a server somewhere, be considered prior art accessible to a POSITA if they are properly indexed?
How Might Brain Interfacing Affect what Qualifies as Prior Art?
The network effect inspires publication of almost everything on the internet, and so it grew as the ultimate repository of documents. With the internet’s growth came changes to the very definition of what qualifies as “prior art”. For example, “written publications,” which are considered prior art, were historically identified as ink documents published by centralized publishing houses. With the internet, what qualifies as prior art expanded to include documents electronically available on the internet. It does not matter whether a document ever was printed to paper. Further, the term includes informal works that may not even be formatted for printing, such as web pages and software code. With the internet, whether a document is considered “publically available” morphed from whether it was copied in sufficient numbers and physically accessible in libraries to whether it was available to the internet and indexed by search engines. The America Invents Act (AIA) even did away with geographic and language restrictions on what defined prior art.
If the commingling of thoughts shared through brain-machine interfaces becomes commonplace, and they become publicly accessible and indexed, then perhaps they may be the next frontier in prior art. That is, if the concept of a “printed publication” expanded from things that are neither printed to paper nor published by a publishing house, then perhaps the concept of prior art will further expand to include electronic representations of thoughts themselves, or perhaps the mixture of different people’s thoughts. At some point, a database of this new prior art of electronic thoughts may be submitted as evidence to defend against patent infringement. It is conceivable that such a database will be proffered by enterprising counsel for an invalidity defense in district court or an inter partes review (IPR).
If that is to be the case, then should an examiner be able to cite stored, indexed brain-machine-interface signals for prior art, or at least as evidence for what is “well known”? Should such art be classified in 35 U.S.C. § 102 any differently from the current three types of prior art (patents, patent publications, and non-patent literature)?
The very fact that electronically stored, commingled thoughts graced different fields in one sitting, or considered a certain number of solutions, conceivably could be used for these purposes. That is, even if the electronically stored thoughts are not considered prior art per se, they may be evidence for combining other prior art or showing something is obvious in another way. Or, they may be submitted by an applicant as evidence of secondary considerations of non-obviousness, such as long-felt but unsolved needs, professional approval, and unexpected results. As applicants are less restricted by rules and training than examiners for what to argue, it may very well be that electronically stored thoughts are proffered first by applicants. Should the United States lead the way in identifying, prohibiting, or regulating their use in patent law?
It is possible that electronic representations of thoughts may not make the cut for the purposes above. However, they may inevitably squeak their way into patent law, such as to prove earlier conception. Although the AIA did away with the importance of proving absolute priority in conception by moving the United States to a “first inventor to file” (35 U.S.C. § 100 (note)) scheme, it replaced it with derivation proceedings under 35 U.S.C. § 135. Electronically recorded thoughts could act as a type of date-stamped laboratory notebook, weighed as evidence by a tribunal deciding whether an applicant derived an invention from another.
How Might Brain Interfacing Affect Disclosure Requirements?
The written description and enablement requirements facilitate ensuring that a patentee possessed a claimed invention and support the quid pro quo of the patent system by providing sufficient disclosure for others to make and use the invention. Currently, these requirements are primarily tied to written words in the specification. However, even now, some other types of disclosures are permitted. For example, MPEP 2402 states: “Where the invention involves a biological material and words alone cannot sufficiently describe how to make and use the invention in a reproducible manner, access to the biological material may be necessary for the satisfaction of the statutory requirements for patentability under 35 U.S.C. 112.”
Suppose that brain-interfacing technology advances to a point where interpretable signals can be regularly collected from large populations of consenting individuals and that a remote data store could store and index representations of the signals. Could the enablement requirement be met (in part or in full) by referencing select signal representations? If the signal representations were tied to “sources” that identified a person from whom the signal originated, could the written description requirement be met (in part or in full) by referencing select signal representations associated with an inventor? How might the answer to these questions depend on whether the signal representations may be downloaded to a person’s brain? And would such a download capability be sufficient to indicate that the invention had been reduced to practice?
Preparing the Patent System
Over the last several years, artificial intelligence (AI), and not brain interfaces, has been the focus of patent offices around the world. They have sought feedback and carefully considered whether and/or how it may be prudent to modify patent requirements and examination procedures based on the increasing power and use of AI. Perhaps a similar reflection will be performed as brain-interfacing technology becomes increasingly powerful and more frequently used.
As noted in the initial paragraphs of this article, a cross-institutional group of authors predicted that this technological future is on the horizon, and DARPA is funding efforts to dramatically enhance technological capabilities of brain interfacing.
Consider: in 1983, the first mobile phone was released, and three decades later, 58% of adults in the United States owned a (much more sophisticated) smart phone. Thus, to the authors, is seems entirely plausible that, over the next 1-3 decades, brain-interfacing technology will undergo major advancements and become staples in our everyday lives. The question, then, is at what point we should prepare our IP systems for this potential future?
Image Source: Deposit Photos
Image ID: 75383115