Mondaq
“Proponents of AI models suggest that copyrighted content is not incorporated verbatim and the content is transformed into something new. But if the content is indeed new, who owns it?”
“Proponents of AI models suggest that copyrighted content is not incorporated verbatim and the content is transformed into something new. But if the content is indeed new, who owns it?”
Choosing the latest emerging technologies with implications for IP is somewhat subjective. I say somewhat because Artificial Intelligence (AI)—and generative AI in particular for purposes of this article—is the elephant in the room. AI is reshaping the world and the practice of law at large. No lawyer or legal professional can ignore AI, even if they ultimately dismiss it. Most are at least exploring AI and considering where it may fit, if anywhere, in their law practice. And unfortunately, some continue making news for AI disasters such as disclosing confidential information and citing nonexistent case law.
Other emerging technologies also present—or continue presenting—notable implications for IP. As expected, dependencies may exist between some emerging technologies enabled by advances to another.
Generative AI
Generative AI has implications for IP, but U.S. patent law and policy has implications for generative AI.
In July, the United States Patent and Trademark Office (USPTO) released revised guidance for patent examiners and practitioners considering whether AI-based claims are eligible for patenting. Under 35 U.S.C. § 101 and prevailing law, claims generally must not be directed to patent eligibility “exceptions”—abstract ideas, laws of nature, or natural phenomena (including products of nature)—or must recite features amounting to significantly more than an exception or use of a computer. Previous guidance set forth generally the multi-step test for patent eligibility and various examples involving application of computers. The revised USPTO guidance adds to the previous guidance discussion and examples of patent eligibility determinations specifically for AI-based claims.
In the courts, at least one statutory interpretation question related to generative AI has been raised at the Federal Circuit. Specifically, whether generative AI may be a named inventor on a patent application. The answer for now is “no”. In other words, generative AI precipitated and provided a legal precedent for whether an “inventor” or “individual” as written in 35 U.S.C. § 100 was intended to include only natural persons.
Despite interesting issues across IP, generative AI’s implications arguably have been felt most in the copyright realm. As one California Judge has noted, analyzing AI tools under the fair use doctrine is still relatively new. The fair use doctrine allows certain uses of copyrighted works without owners’ permission. The judge recognized the issues are “cutting edge” and “test[] the boundaries of fair use and copyright law.” For example, generative AI uses third-party works for training models on which to base outputs. The training itself raises questions regarding whether using third-party works as inputs is a fair use. And the output may infringe copyrights if, e.g., it misappropriates likenesses or existing works. Copyright infringement claims involving AI-generated images and music are proliferating.
Proponents of AI models suggest that copyrighted content is not incorporated verbatim and the content is transformed into something new. But if the content is indeed new, who owns it? U.S. copyright law requires “human authorship” and AI-generated content would apparently not qualify because a human was not the creator. In other words, the content in any event is not the result of human discretion. And the “content” may include not only traditional creative arts but also pervasive commercial works such as marketing campaigns. The issue is receiving attention from both the U.S. Copyright Office and courts.
On the trademark front, whether (and how) generative AI use or output may raise trademark issues—e.g., cause a likelihood of confusion with a trademarked brand—is not so intuitive as, for example, reproduction of a copyrighted work. However, as generative AI becomes more widely used for everything from legal briefing to social media trends, we are learning how it may happen. In one example, social media users using AI tool(s) to produce images of their pets in a style associated with an original content producer inputted the original content producer’s name into the AI prompt. In certain cases, resulting images included the original content producer’s trademarked logo. In a different example, a trademark owner has alleged that generative AI diluted its mark by attaching the mark to low quality, unappealing, or offensive images. The trademark community must be extra vigilant for potential concerns because of the sheer volume of work that generative AI enables.
Trademark owners nonetheless have opportunities for using generative AI to their advantage. For example, generative AI may be used for developing intelligent marketing materials and placements based on the most effective aspects determined through inputs. And what if trademark owners offered licenses to generative AI platforms for deliberately attaching marks to controlled outputs from the AI tool, thus taking advantage of the sheer volume as well?
Interestingly, trademark owners’ use of generative AI towards these advantages may fall within the issues discussed above, such as whether the AI-based content strengthens the mark under the relevant law and who owns the AI-generated content even if it includes a license to the mark.
Quantum Computing
Quantum computing technology uses quantum-mechanical phenomena such as electron states for storing and processing information in quantum bits (“qubits”, pronounced cue-bits). Quantum computers would be a transformative tool in solving complex mathematical problems in boundless contexts.
Accordingly, quantum computing continues to attract much public and private attention and funding. But viable quantum computing will require a host of harmonized developments.
Beyond the extensive hardware and operational tech, the algorithms devised for taking advantage of quantum computing’s versatility may also be protectable IP. For example, the algorithm may represent a patentable advancement in the art. The written code may be protectable under copyright law. In other cases, developers may simply keep their algorithms as trade secrets. And the software and techniques may be unique or proprietary, making trademarks vital in staking claims to the advancements.
Read more at
