Generative AI & Law: Thinking Machines & Patent Law
The evolution of tools has brought us to a unique moment in history—one where machines can "think" in ways previously limited to human cognition. Generative AI, powered by large language models and sophisticated algorithms, now plays an active role in generating novel ideas and solving complex problems. These AI tools (“Thinking machines”) are a radical departure from traditional tools that were passive, requiring humans to supply all the creativity and problem-solving. This paradigm shift presents significant challenges for legal frameworks, particularly patent law, where Section 103 of the U.S. Code lays out the conditions for patentability, including the inventive step. This section was written in a time when it was inconceivable that machines could contribute meaningfully to inventiveness. Now, as AI continues to develop, the law must evolve to address the complexities of human-machine collaboration in innovation.
The Historical Context: Passive Tools of the Past
For most of history, tools were "dumb" objects, designed to assist but never capable of independent thought or creativity. Typewriters, drafting boards, and calculators helped humans execute tasks but did not contribute to the intellectual leap required for invention. Inventors like Thomas Edison and Nikola Tesla used tools to manifest their ideas, but the creative spark—the moment of insight—was undeniably human. Patent law, especially Section 103, reflected this understanding. The inventive step, or "non-obviousness" standard, required that an invention demonstrate a level of ingenuity beyond what was considered obvious to someone skilled in the art. This standard assumed that humans were the sole drivers of innovation, responsible for both framing the problem and devising the solution.
Enter Thinking Machines: A New Era of Innovation
Today, the advent of thinking machines like generative AI has upended this traditional view of invention. These AI systems can analyze vast amounts of data, recognize patterns, and generate outputs that humans may not have considered. In fields such as drug discovery, material science, and even creative industries, AI has demonstrated an ability to suggest novel solutions or designs, often surprising even their human operators.
For example, DeepMind’s AI system, AlphaFold, made groundbreaking advancements in predicting protein structures—an area where traditional human-driven methods had struggled. This leap in biological understanding was facilitated by the AI's ability to analyze vast datasets and recognize patterns that were not obvious to human researchers. In this case, AlphaFold’s contributions highlight the capabilities of thinking machines in problem-solving and invention. But this also raises a critical legal question: Who owns the intellectual property rights to these breakthroughs? If a machine plays a significant role in the discovery, can the resulting invention still be considered a product of human ingenuity?
Inventive Step and AI: Challenges for Patent Law
Section 103 of U.S. patent law is built on the principle of non-obviousness, meaning that an invention must be sufficiently creative or novel, not something that would be easily deduced by someone skilled in the field. This requirement was crafted at a time when humans were the only creative agents. Thinking machines complicate this principle because they are capable of performing tasks that once relied solely on human creativity.
One of the key challenges is understanding how AI fits into the inventive step requirement. Traditionally, the inventive step was entirely the domain of the human inventor, with the tool playing a purely supportive role. However, with AI-driven innovation, the line between tool and creator blurs. AI can frame problems in creative ways, suggesting entirely new approaches or combinations of elements that might not have been obvious to a human. For example, generative AI systems have been used in the automotive and aerospace industries to design new structures and materials, resulting in inventions that surpass traditional methods.
This leads to the question of how patent law should account for AI's contribution. If an AI is used to frame the problem or generate solutions, does that diminish the human inventor's role in meeting the non-obviousness standard? In practice, should the inventive step be credited to the human who directed the AI, or does the AI’s role necessitate a rethinking of the way patents are granted?
Creative Problem Framing: A New Frontier
One of the most striking ways thinking machines are transforming invention is through what might be called "creative problem framing." In traditional patent law, inventors are praised not only for solving a problem but also for recognizing that the problem exists in the first place. AI systems, however, are now capable of identifying patterns and opportunities that humans may overlook, reframing the problem in a way that leads to novel solutions.
For example, in architecture and engineering, a generative AI tool (a ‘machine’) can analyze thousands of potential designs to optimize for factors like cost, material efficiency, and aesthetics. While the human operator sets the initial parameters, the AI's ability to generate and evaluate countless possibilities often results in solutions that are both innovative and unexpected. This raises the question: Is the human responsible for the inventive step, or should the AI’s role in framing and solving the problem be acknowledged in the patenting process?
Where to "Aim" the AI Tool: Shifting the Role of Human Inventors
With thinking machines, the human inventor’s role often shifts from direct problem-solving to guiding and directing the AI toward a particular outcome. In this context, the inventor is not necessarily creating the solution, but rather setting the AI on a path toward discovery. This dynamic creates an interesting question for patent law: Does the act of aiming the AI tool represent a sufficient inventive step? Is the human’s role in choosing where to direct the AI’s capabilities enough to satisfy the non-obviousness standard, or does the AI’s contribution overshadow that of the human?
Consider the use of AI in pharmaceutical research. Scientists might direct an AI system to analyze a vast dataset of chemical compounds to identify potential candidates for a new drug. While the human sets the goal, the AI does much of the heavy lifting, identifying relationships and patterns that lead to the discovery of a novel compound. In this case, the AI’s ability to sort through data and recognize potential leads raises questions about how much credit the human should receive for the inventive step.
Recognizing Useful Outputs: Human vs. Machine Creativity
Another critical issue is how to recognize when AI generates a promising or useful output. Traditionally, human inventors were responsible for recognizing when they had hit upon a valuable invention. With AI, however, the system itself can often flag valuable results. This shifts the dynamic further away from the human and more toward the machine as an active contributor to the inventive process.
For example, OpenAI’s GPT models have been used to generate novel designs in fields like fashion and industrial design. These models can produce creative outputs that a human operator might not have considered. If an AI generates an inventive design and then flags it as useful or novel, who should receive credit for the discovery? The human who initiated the process or the machine that produced the result?
A Future of Hybrid Innovation: The Need for Legal Evolution
As thinking machines continue to advance, it is clear that patent law, and particularly Section 103, must evolve to address these new complexities. The traditional view of tools as passive and inventors as the sole creative agents no longer fits the current reality. Patent law must begin to account for the active role that AI plays in framing problems, generating solutions, and recognizing valuable results.
One potential approach could involve redefining the concept of inventorship to allow for hybrid human-AI inventors. This would acknowledge the unique contributions of both the human and the machine in the inventive process with the work done by a machine being considered as work-product. Another option might involve creating new legal frameworks that differentiate between inventions generated solely by humans and those where AI played a significant role.
Regardless of the solution, it is clear that thinking machines are here to stay. Generative AI’s ability to engage in creative problem-solving challenges the foundational principles of patent law, and the legal system must adapt to ensure that inventors, whether human or machine, are given appropriate credit for their contributions to innovation.
Further read
From Infinite Improbability to Generative AI: Navigating Imagination in Fiction and Technology
Human vs. AI in Reinforcement Learning through Human Feedback
Generative AI for Law: The Agile Legal Business Model for Law Firms
Generative AI for Law: From Harvard Law School to the Modern JD
Unjust Law is Itself a Species of Violence: Oversight vs. Regulating AI
Generative AI for Law: Technological Competence of a Judge & Prosecutor
Law is Not Logic: The Exponential Dilemma in Generative AI Governance
Generative AI & Law: I Am an American Day in Central Park, 1944
Generative AI & Law: Title 35 in 2024++ with Non-human Inventors
Generative AI & Law: Similarity Between AI and Mice as a Means to Invent
Generative AI & Law: The Evolving Role of Judges in the Federal Judiciary in the Age of AI
Embedding Cultural Value of a Society into Large Language Models (LLMs)
Lessons in Leadership: The Fall of the Roman Republic and the Rise of Julius Caesar
Justice Sotomayor on Consequence of a Procedure or Substance
From France to the EU: A Test-and-Expand Approach to EU AI Regulation
Beyond Human: Envisioning Unique Forms of Consciousness in AI
Protoconsciousness in AGI: Pathways to Artificial Consciousness
Artificial Consciousness as a Way to Mitigate AI Existential Risk
Human Memory & LLM Efficiency: Optimized Learning through Temporal Memory
Adaptive Minds and Efficient Machines: Brain vs. Transformer Attention Systems
Self-aware LLMs Inspired by Metacognition as a Step Towards AGI
The Balance of Laws with Considerations of Fairness, Equity, and Ethics
AI Recommender Systems and First-Party vs. Third-Party Speech
Building Products that Survive the Times at Robometrics® Machines
Autoregressive LLMs and the Limits of the Law of Accelerated Returns
The Power of Branding and Perception: McDonald’s as a Case Study
Monopoly of Minds: Ensnared in the AI Company's Dystopian Web
Generative Native World: Digital Data as the New Ankle Monitor