Can Generative AI Further Enable the Promise of Patents?
Throughout my thirty-plus years as an IP lawyer, I have had a lot of experience with transactions and disputes involving patents. I wondered how transaction velocity[1] could be significantly improved and how litigation could be minimized.[2] I have found that transaction inefficiencies are typically caused by differences in access to information about patents, causing disputes and litigation between parties. These inefficiencies also detract from the patent system’s promise of providing innovation incentives.[3] I believe that these failings will ultimately be resolved by generative AI tools that enable better information symmetry for promoting patent transactions, thus helping to resolve patent disputes and fulfilling the patent system’s promise of promoting innovation incentives.
Pitfalls with Today’s Patent Dispute Resolution Process
Patent documents are typically long, technically complex and nuanced, leaving them open to interpretation. This leads to frequent disagreements between parties, as it can be unclear whether a patent infringes or is valid. When this occurs, parties typically either negotiate (leading to a license or some other compromise), or they litigate, in which case the courts decide the issues for them. In my experience, enduring the process of protracted patent litigation was the norm rather than entering efficient settlements or license transactions and potentially avoiding litigation altogether.
One major reason negotiations break down is asymmetry in information between the parties. The parties involved in negotiations usually must hire legal and technical experts to determine the patents’ true scope and strength. Not only is the cost inefficient, but it also leads to the parties having sharply differing opinions during negotiations because their respective experts seldom arrive at the same conclusions. Thus, negotiations typically fail, and patent litigation occurs. Eventually, the court provides the parties with symmetrical information by determining the scope of the patent, validity[4], and infringement determinations, and this can enable the parties to eventually settle (a form of transaction) and avoid the high costs incurred to maintain the litigation as an incentive to reach a deal.
The high transaction costs associated with patent dispute resolution—legal fees, expert witnesses, lengthy discovery processes—along with non-monetary costs such as distraction and delay, create opportunities for inefficient outcomes. These costs not only affect the concerned parties, but also place a substantial burden on societal resources for resolving disputes (e.g. judges and jurors)[5]. These factors disincentivize further innovation by diminishing the promise of patents to incentivize, thus reducing their ability to foster innovation that benefits society. The cost inefficiency also extends to opportunistic behaviors such as so-called “efficient infringement” for by accused infringer or a patent holder’s practice of so-called “patent hold-up.” These parties leverage the high litigation costs and time delays to extract opportunistic transactions and settlements. For efficient infringement, which is the practice of litigating against an organization with significantly fewer resources, the smaller organization is forced to settle because they lack the resources to maintain the litigation. Occasionally, patent holders practice patent hold-up, which is when a patent holder asserts patent infringement (usually on a weak patent), resulting in the other party settling because litigating the case would be more expensive. If these costs and delays could be substantially reduced, transaction efficiency for patents would be substantially improved.
Avoiding Pitfalls by Using Generative AI Tools
Ronald H. Coase, considered by many to be the pioneer of the “law and economics” field, suggested that under ideal conditions—well-defined property rights and symmetrically available information—parties to a dispute can achieve optimal efficiency in transactions through their negotiations (also called the Coase Theorem). However, real-world conditions rarely meet these standards, particularly in the complex realm of property rights, e.g. patents, forcing the parties to resolve their disputes in court.[6] Parties and courts suffer from information asymmetry since they are not equipped to efficiently assess the technical complexity, claim language and other legal nuances of patents and patent law.
Generative AI has the potential to help reduce this information asymmetry and increase transaction velocity, at every level and step in the legal process. With capabilities such as advanced text generation, detailed text and figure analysis, data synthesis, and more, AI tools can provide an unbiased, clearer, more accurate, and cost-effective understanding of a patent’s scope, infringement and validity strengths and weaknesses.
One of GenerativeIQ LLC’s portfolio companies, Patlytics (www.patlytics.ai), offers a generative AI solution for litigation support and efficient patent evaluation by assessing a patent’s scope, validity, and infringement. Specifically, Patlytics’ patent litigation platform performs:
Automated Patent Analysis: Patlytics can meticulously analyze patent claims, specifications, prosecution history, and relevant technical literature, offering balanced insights into the patent’s claim interpretation, validity and scope.
Cost-Effective Technical Analysis: Patlytics can automatically conduct technical discovery generating detailed claim charts for validity and infringement purposes that would traditionally require expensive expert interventions and countless attorney hours.
Symmetrical Information: By generating confidential, detailed, and unbiased analysis, Patlytics’ goal is to ensure that parties involved in patent conflict resolution can have cost-effective access to critical information, helping to minimize gaps and discrepancies in understanding, facilitating negotiations, transactions, settlements and more efficient litigation.
Conclusion
While real-world conditions may never perfectly meet the ideals described by Coase’s Theorem, generative AI tools, such as Patlytics, represent a significant step towards improving the efficiency of the patent eco-system. They substantially reduce the costs of obtaining reliable information, making it more symmetric for all parties involved. By providing better information and reducing transaction and litigation costs, these tools help in fulfilling the patent system’s promise of incentivizing innovation and ensuring societal benefits from technological advancements.
[1] Speed for 1) determining whether there is infringement of a patent, 2) assessing validity of the patent and 3) negotiating a transaction for the patent (e.g. license, purchase, financing, etc.). These three factors are the major factors that go to the market efficiency of transaction velocity for patents.
[2] Including friction-costs for legal fees, expert witnesses, lengthy discovery, trials, and appeals —along with non-monetary costs such as distraction and delay.
[3] The basic goal of the patent system is to encourage innovation. By giving patent-holders the exclusive right to make, use and sell their patented technology, patent-holders are economically incentivized to innovate.
[4] Patent Trial and Appeal Board (PTAB) is a tribunal that reviews issued patents from the U.S. patent office with a specialized panel of administrative judges, who are typically lawyers with technical backgrounds. The PTAB provides an alternative forum to district courts for deciding the validity of patents. In 2011, Congress promulgated the PTAB to help more efficiently decide the validity of issued patents. The party sued in district court for patent infringement (the defendant) often files a petition at the PTAB to challenge the validity of the patent. While the PTAB has helped to reduce the informational gap between the parties (concerning the validity of patents), it has not substantially reduced litigation costs associated with patent disputes.
[5] Courts deciding these issues have lay judges and jurors which face significant challenges understanding the technology and legal issues. Due to the high costs and informational gaps between the parties and the court, resolution of patent conflicts is a highly inefficient process for all involved.
[6] Ron H. Coase, “The Problem of Social Cost,” The Journal of Law and Economics, University of Business and Law Schools Vol. 3, Oct. 1960. Coase received the Nobel Prize in Economics in 1991.
