Artificial intelligence dominated this year's emerging technology updates from the U.S. Patent and Trademark Office and the U.S. Copyright Office. These agencies, among many others, were kept quite busy under the directives set forth in President Joe Biden's Oct. 30, 2023, executive order on the safe, secure, and trustworthy development and use of AI.
The executive order directed the USPTO to (1) publish guidance addressing inventorship and the use of AI in the inventorship process, and (2) issue additional guidance to address other considerations at the intersection of AI and intellectual property, which could include updated guidance on patent eligibility.
The Copyright Office has also been busy working on a multipart report on the effect of AI on copyright law, with Part 1 of the report published in July recommending legislation to create a right to control digital replicas of one's voice and likeness, and the announcement of several Digital Millennium Copyright Act exemptions — but not for AI research.
Additionally, on the trademark front, the USPTO has changed its audit procedure to weed out digitally generated specimens.
Set forth below are highlights on procedural and substantive IP issues in 2024 from the Copyright Office and USPTO, followed by what to expect in 2025.
Patent Updates
On Feb. 13, the USPTO issued inventorship guidance for AI-assisted inventions, applicable to all applications.[1] The inventorship guidance reiterates that all inventors must be natural persons, citing the 2022 case of Thaler v. Vidal in the U.S. Court of Appeals for the Federal Circuit.
The inventorship guidance also sets forth that, for AI-assisted inventions, each named natural person inventor "must have made a 'significant contribution' to the claimed invention," citing the factors set forth in the Federal Circuit's 1998 decision in Pannu v. Iolab Corp. The inventorship guidance further reminds practitioners of their duty of disclosure and duty of reasonable inquiry as they pertain to inventorship.
On April 11, the USPTO issued guidance on the use of AI-based tools in practice before the USPTO.[2] The guidance notes the increasing use of AI-based tools, from examiners performing AI-enabled prior art searches, to practitioners using AI-based tools to research prior art, automate the patent application review process and gain insight into examiner behavior.
The guidance highlights the USPTO's existing rules, including, inter alia: the duty of candor and good faith, e.g., duty of disclosure of AI prior art searches; the signature requirement, or certification of human review of AI-generated content; confidentiality, or risk of disclosure of confidential information through third-party AI systems; and foreign filing license, or improper export of information through AI systems.
On April 30, the USPTO published a request for comments regarding the effect of the proliferation of AI on prior art, the knowledge of a person having ordinary skill in the art, and determinations of patentability made in view of the foregoing.[3]
The request for comments raises many issues posed by AI systems, including the proliferation of AI-generated disclosures, e.g., the "All Prior Art" project; qualifications of a printed publication generated solely using AI that was never reviewed by a human; and the relevant knowledge and level of skill of a person having ordinary skill in the art; among other issues.
On July 17, the USPTO issued a guidance update on patent subject matter eligibility, including on AI.[4] The guidance update focuses on the evaluation of whether a claim recites an abstract idea (Step 2A, Prong 1) and the evaluation of the improvements consideration (Step 2A, Prong 2.)
The guidance includes new Examples 47-49 addressing the use of an artificial neural network to identify or detect anomalies, AI-based methods of analyzing speech signals and separating desired speech from extraneous or background speech, and an AI model designed to assist in personalizing medical treatment to the individual characteristics of a particular patient.[5]
Copyright Updates
Copyright Office urged Congress to address generative AI-generated digital replicas.
In July, the Copyright Office issued a recommendation urging Congress to protect the right of publicity of individuals in view of the increasing "speed, precision, and scale of AI-created digital replicas" of individuals.[6]
The recommendations from the Copyright Office came on the heels of growing concern regarding the sophistication and wide availability of generative AI tools to mimic the sound and appearance of people. Digital replicas are generally considered to be technologies that can replicate people's voices and appearance to the point where it is difficult to tell what's real and what's the replica.
As a prime example, the Copyright Office pointed to a robocall replicating Biden's voice that discouraged people from voting in the New Hampshire primary earlier in 2024. To help stem this growing problem, the director of the Copyright Office and register of copyrights stated that the office believes "there is an urgent need for effective nationwide protection against the harms that can be caused to reputations and livelihoods," and urged Congress to consider this recommendation.
Current laws such as the Lanham Act — which has provisions against false association — are much too narrow to adequately address digital replicas, according to the Copyright Office, since those laws would only apply in commercial circumstances.
While the Copyright Office did note some benefits to the availability of digital replicas in circumstances in which an artist is no longer available to perform, it nevertheless concluded that the "broad range of actual or potential harms arising from unauthorized digital replicas" necessitates congressional action given one of the highlighted harms related to the creation of explicit content.
Indeed, the recommendation highlighted a study's findings that the vast majority of deepfake videos online were pornographic in nature and targeted women. This danger is in addition to the integrity of the political arena by the fake robocall replicating Biden's voice that was cited.
In response to the recommendations from the Copyright Office's concerns, Congress proposed the Nurture Originals, Foster Art, and Keep Entertainment Safe, or NO FAKES, Act in October. Under the NO FAKES Act, there would be liability for creating, hosting or sharing an unapproved digital replica, and online service providers would have to remove such content after receiving notice. The act does have First Amendment exclusions for protected expression such as documentaries or parodies.
The NO FAKES Act does preempt state law addressing digital replicas in its current form, which will be an interesting area to monitor since the Copyright Office urged no preemption for the sake of clarity.
Copyright Office denies DMCA exemption for AI research.
In October, the U.S. Copyright Office announced several Digital Millennium Copyright Act exceptions for categories such as data mining for scholarly research. However, the Copyright Office denied a requested exemption for acts relating to the trustworthiness research for generative AI platforms.
The petition was filed in 2023 by Jonathan Weiss of Chinnu Inc., an information technology security and consulting company, and sought an exemption under Title 17 of the U.S. Code, Section 1201, for circumvention of access controls on generative AI platforms for the purpose of researching biases in the platforms. If approved, the exemption would allow the sharing of research for highlighting bias in generative AI models that has been flagged as a concern.
Nevertheless, the Copyright Office ruled against the proposed exemption on the basis that congressional action to amend the law was the proper means for addressing issues relating to the trustworthiness of generative AI platforms.
In explaining the denial, the associate register of copyrights noted that the cause of adverse effects in generative AI platforms was due to the large number of rules that the model providers had in place in terms of service and safety guidelines. The Copyright Office therefore reasoned that an exemption would not actually solve the identified problem.
Trademark Updates: Changes to Audit Procedure
The USPTO is changing its audit procedures of maintenance filings in its ongoing battle against digitally generated, including generative, AI and other types of specimens of use that the USPTO suspects to be false.
The new procedure will allow direct auditing of filings to challenge suspicious specimens rather than just relying on the random audits that have been in place since 2017. The change is necessary due to increasing instances of digitally created specimens that allow registrations to falsely remain in effect even though there is no actual use in commerce, which circumvents use requirements for maintaining a registration.
To remain active, all U.S. trademark registration must file a declaration of continued use, or excusable nonuse, between the fifth and sixth year of registration, and every 10th anniversary, to remain pursuant to Section 8 of the Trademark Act or Section 71 for International Registrations, or the Madrid Protocol.
The declaration of use must be accompanied by a specimen showing the current use of the mark with the goods and services in the registration. The purpose of this requirement is to keep the trademark registry clear of "deadwood" so that new filings are not barred by older registrations that are not actually in use.
In 2017, the USPTO implemented a policy to randomly audit specimens filed with declarations of continued use. The goal was to weed out registrations where there was not actual use at all or the mark was not in use for all of the goods and services in the registration.[7]
The procedure was that an office action would be issued requiring further proof of use. The audit procedure was effective as there were financial penalties if a response requested deletion of goods or services where there was no evidence of use. The audit system also uncovered "systemic efforts to subvert" the use-in-commerce requirement to maintain registrations that should have been rightfully canceled.
However, the 2017 audit policy was not able to address the recent influx of digitally generated specimens from online "specimen farms" and user-generated generative AI specimens. Also, the random audit policy did not allow the USPTO to directly challenge a suspicious specimen.
The USPTO stated that it recently became increasingly aware of digitally created specimens generated to subvert the use in commerce requirements to maintain registrations, such as the use of specimen farms and generative AI platforms which create the illusion of commerce without providing actual sales.
The new rules will direct audits of suspicious filings that show obvious signs of digital alteration or specimen farm website use. The objective of the directed audit program is "to promote the accuracy and integrity" of the trademark register. If the recipient of an audit is unable to prove an actual use of the mark, the registrations will be canceled.
The revised audit policy will accomplish the goal of helping avoid the "needless costs and burdens" to others in adopting new marks unnecessarily and avoid conflicts with a registered mark when it is not actually in use.
As outgoing USPTO Director Kathi Vidal correctly stated in a USPTO trademark alert published in October, "conducting post registration audits is just one of the ways we are ensuring the integrity of the trademark register for the benefit of all."[8]
A Look Toward 2025
Many of the foregoing updates were initiated under Biden's executive order, and it remains to be seen whether President-elect Donald Trump will follow these or other AI initiatives in his second term, or what the policy directives for the new USPTO and Copyright Office leadership will look like. However, early predictions are that the new administration will seek to take a more hands-off approach regarding AI regulation, but the actual initiatives remain to be seen.
Nevertheless, the rapid development and widespread adoption of AI will almost certainly continue to disrupt settled areas of IP law and practice in 2025, and agencies, courts and legislators will need to continue to engage with stakeholders as novel substantive and procedural issues emerge. It is clear that the relevant IP agencies are attempting to adapt and react to the new AI world.
This article was originally published in Law360's Expert Analysis section on January 9, 2025. Read more at: https://www.law360.com/articles/2281493/reviewing-2024-s-ai-patent-and-copyright-developments.
[5] https://www.uspto.gov/sites/default/files/documents/2024-AI-SMEUpdateExamples47-49.pdf.
[8] https://www.uspto.gov/subscription-center/2024/changes-how-uspto-selects-registrations-audit.
