“IP law is likely to attract the attention of the next administration and Congress, given the rapid rate of technological advancement and related uncertainties.”
The debates surrounding the 2024 election focused on “hot button” issues like abortion, immigration, and transgender rights. But several important IP issues also loom over the next administration and Congress. These issues include AI-generated deepfakes, the use of copyrighted works for AI training, the patentability of AI-assisted inventions, and patent subject matter eligibility more generally. We might see President Trump and the 119th Congress tackle some or all of these issues in the next term.
AI-Related Issues
AI is rapidly growing and, for better or for worse, becoming a part of everyday life. Students rely on ChatGPT for their coursework, deepfakes impersonate politicians, and molecular biologists employ deep?learning algorithms to advance drug discovery and enhance medical diagnoses. The growing reliance on and development of AI has raised interesting issues in both the copyright and patent arenas. Possible legislation and executive initiatives might address deepfakes, the use of copyrighted works in training AI models and generating content from those models, and the patent eligibility of AI-assisted inventions.
AI-Generated Deepfakes
Deepfakes are a likely target for new legislation over the next four years. The Copyright Office published a report in July 2024 calling for prompt federal action to regulate deepfakes, which the Copyright Office defines as video, image, or audio recordings digitally created or altered to realistically but falsely depict individuals. The Copyright Office notes that deepfakes may threaten the entertainment industry, political processes, and the livelihoods of private individuals who were impersonated.
There are already multiple congressional proposals targeting deepfakes, which range from regulating deepfakes in specific contexts, such as political campaigning, to possibly creating whole sets of new intellectual property rights. For example, the REAL Political Advertisements Act would require political advertisements to disclaim the use of AI-generated sounds or images. Meanwhile, the No AI Fraud Act would more broadly establish intellectual property rights in voice and likeness, ultimately protecting against the use of unauthorized deepfakes that readily identify an individual. It will be interesting to see the extent to which the new administration pursues regulation of deepfakes.
AI Training
The new administration and Congress may also seek to regulate how AI models are trained and decide whether the use of copyrighted materials in AI training constitutes infringement or fair use. Plaintiffs in several pending lawsuits allege that AI models were unlawfully trained on copyrighted works, which AI developers often obtain from web scraping. AI developers argue that the use of copyrighted works to train AI models is transformative fair use. As these issues make their way through the judicial system, the new administration or Congress could intervene with proposed legislation to address them.
Relatedly, plaintiffs in these lawsuits often struggle to identify in their complaint their allegedly infringed works due to lack of visibility into how AI models were trained. Congressman Adam Schiff (D-CA) has introduced the Generative AI Copyright Disclosure Act, which would require public disclosure of the materials that developers use to train their AI models. If passed, this Act could help plaintiffs build their case against AI developers.
AI-Assisted Inventions
The United States Patent and Trademark Office (USPTO) issued guidance in February 2024 on the patentability of AI-assisted inventions, emphasizing that the inventorship analysis should focus on whether there was “significant contribution” by a human being. According to the USPTO guidance, in accordance with the Federal Circuit’s decision in Thaler v. Vidal (2022), an AI system cannot be listed as an inventor, as an “inventor” under 35 U.S.C. § 100(f) must be human. But a person’s use of AI to create an “AI-assisted invention” does not prevent that person from obtaining a patent, so long as that person “contributes significantly” to the invention.
New legislation might seek to depart from the USPTO guidance and the Federal Circuit. Such legislation could articulate the minimum extent to which a human being must contribute towards the inventive process or could redefine “inventor” in 35 U.S.C. § 100(f).
Patent Subject Matter Eligibility and 35 U.S.C. § 101
Other possible legislation during the next administration may seek to reform 35 U.S.C. § 101, which addresses the subject matter eligible for patent protection. Much ink has been spilled on the Supreme Court’s landmark decisions interpreting Section 101. The new president and Congress may look to clarify or change the bounds of subject matter eligibility with new legislation.
By way of background, the Supreme Court’s test for determining whether patent claims are eligible subject matter—first conceived in Mayo Collaborative Services v. Prometheus Laboratories, Inc. (2012) and expanded on in Alice Corp. v. CLS Bank Intl’l (2014)—unfolds in two inquiries. The first inquiry is whether the claims are directed to patent-ineligible subject matter, such as laws of nature, natural processes, and “abstract concepts.” If so, the second inquiry looks at whether the claims add an unconventional, “inventive concept” that makes the otherwise ineligible claims become eligible.
Many practitioners and academics have argued that this two-part test is difficult to apply and that lower courts have strained to apply it consistently and clearly. What constitutes an abstract concept? What is sufficiently inventive? What’s the difference between Section 101 inventiveness and Section 103 non-obviousness? These are but a few of the prominent questions one might hear echoing in the halls of intellectual property law firms and the office hours of patent law scholars.
The complaints have already received attention during recent congressional sessions, and those frustrated will want to track the developments of proposed legislation. One such proposal is the Patent Eligibility Restoration Act, which was introduced in 2023 by Senators Thom Tillis (R-NC) and Christopher Coons (D-DE). Of its many interesting ambitions, the bill seeks to mandate that eligibility shall be determined without considering, among other things, whether a claim element is “known, conventional, routine, or naturally occurring.” If enacted, the Act would thus seem to abrogate the Mayo test and much of the jurisprudence on Section 101 over the past 12 years. Most recently, Tillis withdrew the bill from consideration by the full Senate Judiciary Committee, citing forces at work that “may be trying to undermine progress of either PERA or PREVAIL,” but said that the bill will move forward regardless, “so those [trying to thwart the bills] need to get at the table before they get on the table.”
Senators Tillis and Coons proposed the bill with hopes that it would bring “vital clarity for inventors and innovators.” Only time will tell whether confusion surrounding Section 101 will actually clear, or if the Act will conjure uncertainty in a new form.
Final Remarks
IP law is likely to attract the attention of the next administration and Congress, given the rapid rate of technological advancement and related uncertainties in IP law, such as those discussed above. New IP legislation and regulation would certainly seize the attention of practitioners and the industries they serve. Given the importance of these matters to the global economy and (in the case of deepfakes) democracy, they might even become “hot button” issues for the general public in upcoming years.
