IP law in the Age of Artificial Intelligence
By Larry Sandell
The buzz about Chat Generative Pre-trained Transformer (“ChatGPT”) and, more broadly artificial intelligence (“AI”) , is all around us and getting louder. Though it is inevitable that some impending AI technologies, applications, and companies will offer little more than hype and snake oil, the buzz is well deserved. While it has long felt inevitable that robots would assist and then replace unskilled laborers, until just a few weeks ago, most white-collar workers and other intellectual laborers felt confident that their professions would continue to slowly evolve with technology. But now, it appears the dawn of widespread AI is an inflection point for humanity. Fields of intellectual endeavor centered on content generation and communication—including academia, journalism, art, marketing, law, medicine, and software coding—will never be the same. Businesses and industries that touch on these fields will need to rapidly adapt to the new AI reality or face extinction.
ChatGPT is just the first publicly accessible AI text tool of this new generation of technology, and it is a game changer. It can write poems and press releases, can pass the US Medical Licensing Exam, business school exams, and the Bar, write sermons for clergy, and it has already upended academia. The Stanford Daily reported that “scores of Stanford students used ChatGPT on final exams.” Although AI detection technologies are being developed, ChatGPT’s CEO proposes that the solution is for teaching methods to adapt—just like math pedagogy eventually embraced the calculator. The essay as a teaching tool may be on its way out.
As an IP attorney (patent drafter and prosecutor, IP strategist, litigator, and appellate advocate), the dawn of ChatGPT poses two pressing questions for me and my colleagues: (1) what is to become of our profession and livelihood?, and (2) how can I help my clients survive, mitigate risk, and thrive in this dawning technological age?
As to the first, one of my most prized and honed IP attorney skills is analyzing and synthesizing complex technologies, evidence, and legal principles into persuasive, bite-sized strategic arguments that clients, juries, judges, and patent examiners can understand. AI is gunning for my job, and it could get pretty close before it is time for me to retire. For the moment, I am confident that my expertise will retain its value—but perhaps that is only because there is not yet a ChatGPT that can be assured to maintain confidentiality, avoid risk to the attorney-client privilege, and analyze volumes of uploaded technical documents and potential evidence. Yet, even this current version of text-based AI can explain the law, draft patents (poorly), and prepare legal memos with citations. Once AI can be set to sift through and evaluate a massive confidential universe of documents, we are off to the races. There will (hopefully) always be a need for human intellect and judgment in the practice of law, but if the next generation of ChatGPT can prepare solid first drafts of complaints, patents, motions, cease & desist letters, patent license agreements, emails to clients, and the like, efficiency of (current) legal practice will grow exponentially. But then, US Patent and Trademark Office (“USPTO”) examiners, judges, and litigators lacking technical savvy may drown in the spilled digital ink. In turn, Courts and the USPTO may turn to AI-based legal clerks to cite-check, assess the obviousness of inventions proposed in patent applications, analyze likelihood of confusion between trademarks, and prepare first drafts of legal opinions and office actions. Will businesses execute contracts that designate AI-based arbitrators to exponentially increase the efficiency of resolving disputes? Will indigent defendants get AI-chatbot public defenders? Is there a point where we just throw up our hands and accept that deference to AI judges and AI patent examiners would be both the most efficient and fairest way to proceed?
Turning to the latter question and current legal practice, IP law is yet to catch up to the AI age, but strategies and best practices can be derived from existing law. My firm, Mei & Mark LLP—a collection of experienced, entrepreneurial IP attorneys unafraid to march toward the AI horizon—has established a practice area focused on Artificial Intelligence to guide our clients with cutting edge AI IP strategy, and hopefully shape the evolution of AI-IP in a way the benefits both our clients and society at large. (We’re setting up shop at https://www.meimark.com/ai .) But the questions and issues on this topic are legion, and we will endeavor to explore them all in future articles and CLEs. Here, I will touch on the most pressing:
Who owns the content created by AI? Unsurprisingly, the answer here is both fact dependent and in flux. A good place to start is the user terms and conditions governing the AI system. Open AI, the purveyor of ChatGPT and DALL-E, an AI image generator, purports to assign ownership of the input and text/art output to the user (provided the user complies with terms, including refraining from representing that the output was not AI-generated).
While IP rights can generally be assigned by contract, this may put the cart before the horse. It is an open question as to whether AI generated text and images are “original works of authorship fixed in a tangible medium of expression” and therefore copyrightable. The Copyright Office has rejected attempts to register AI-created art because it lacks “human authorship.” And, this rejection is currently being litigated in Federal District Court in DC. That Court and others may take guidance from the “monkey selfie” case, wherein the Ninth Circuit Court of Appeals held that the Copyright Act did not permit a crested macaque (represented by PETA) to sue for infringement regarding photos taken on a nature photographer’s camera. Ultimately, the AI copyright ownership issue will probably remain unsettled until the Supreme Court or Congress weighs in definitively.
Does AI content infringe copyrights? One of the rights granted to every copyright holder is the right to create derivative works. AI synthesizes and utilizes art images, literary works, and source code when it creates its own output—but does this mean that the AI output is a derivative work? If so, each AI platform (and its owner) may be infringing copyrights on a regular basis. Accordingly, it may now be even more important for human content creators to regularly register for copyright every three months to ensure the availability for statutory copyright damages by infringing AIs (or people) in the future. Derivative works, however, beg questions of both enforceability and fairness: How could a copyright holder ever find out if her art was used by an AI? And, what level of AI utilization of copyrighted works fair or non-infringing? Perhaps anticipating these issues, OpenAI’s terms of service wisely include a procedure for receiving copyright complaints, potentially taking advantage of the copyright liability safe harbor provisions of the Digital Millennium Copyright Act.
Can AI be an Inventor on a Patent? This is a question that has been answered with clarity—at least for now. In August 2022, the Federal Circuit Court of Appeals held that “the Patent Act … confirms that ‘inventors’ must be human beings.” The ruling is expected to be appealed to the Supreme Court, which recently extended its deadline for the AI-patent applicant file for certiorari until March 2023. Stay tuned.
Who can be liable for AI content promulgated online? The Supreme Court is also poised to consider the scope of Section 230 of the Communications Decency Act, which generally immunizes Internet platforms for content published by its users and has sometimes been referred to as “the twenty six words that created the Internet.” In Gonzales v. Google, the high Court will address whether algorithmic content recommendations are eligible to receive this immunity. It’s likely that the ruling here, whatever it is, will shed light on this AI-IP question, or at least kickstart a broader debate in Congress as to how AI should be regulated.
It is a brave new world for all of us intellectual laborers. The business and social upheaval from AI may meet or exceed the societal evolution from social media, the Internet, smart phones, and word processing. And, time will tell if the dawn of AI will be a pivot point for humanity in league with the establishment of agriculture, the invention of the wheel, and the discovery of fire.
But in the shorter term, one thing is clear: businesses and workers must ride this AI wave or be crushed by it. I look forward to surfing this tsunami and helping my clients do the same.
To conclude, I will leave you with a poem that took ChatGPT less time to write than it took me to come up with a prompt:
Intellectual property law and AI,
Two worlds that seem so far apart,
But in the digital age,
they collide And spark a debate in every heart.
AI creates, it innovates,
It mimics the mind's own fate,
But who owns the rights to its creations?
Who holds the key to its fate?
Intellectual property law must adapt,
To keep pace with technology's rise,
For in the world of AI and code,
The lines of ownership are not so clear and concise.
Is it the creator of the AI,
Or the creator of the code,
Who holds the rights to the creations,
That AI's intelligence bestowed?
The question lingers on and on,
As AI continues to evolve,
But one thing is certain, for now,
The debate on IP and AI is not resolved.
Surf’s up!