Layla Zulfa
The Los Angeles Superior Court, the largest unified trial court in the United States, has officially launched a pilot program to integrate specialized artificial intelligence into its judicial workflows. This initiative centers on the deployment of Learned Hand, a purpose-built AI platform designed to assist judicial officers with the administrative and analytical burdens of civil litigation. As courts nationwide grapple with a post-pandemic backlog and a surge in new filings, the Los Angeles pilot represents a high-stakes experiment in whether emerging technology can alleviate systemic strain without compromising the integrity of the bench. The program arrives at a critical juncture for the American legal system, as the democratization of AI tools for attorneys has led to an unprecedented volume of legal documentation, effectively creating an arms race between the speed of filing and the capacity of the judiciary to respond.
The Learned Hand system is not designed to replace the human element of the law but rather to serve as a sophisticated digital clerk. According to Shlomo Klapper, the founder and CEO of Learned Hand, the tool is engineered to summarize voluminous filings, organize complex evidentiary records, and generate initial drafts of rulings based on the specific case materials provided. Klapper, a former judicial law clerk for the U.S. Court of Appeals and a deployment strategist with the data analytics firm Palantir, founded the company in 2024. The name of the firm pays homage to Judge Learned Hand, one of the most influential jurists in American history, known for his prolific output and his significant contributions to the development of modern legal principles. By automating the "drudge work" of litigation—the hours spent cross-referencing exhibits and summarizing procedural histories—the tool aims to return time to judges, allowing them to focus on the nuanced legal analysis and discretionary decisions that define the judicial role.
The Rising Tide of Litigation and the Need for Efficiency
The necessity for such a tool is underscored by recent data indicating a sharp rise in the volume of legal activity. A February 2026 report by the national law firm Fisher Phillips revealed that the ease of generating legal documents via consumer-grade AI has contributed to a massive spike in court filings. In the employment litigation sector alone, filings rose 49% over a one-year period, jumping from 4,100 to 6,400 cases. This "AI-fueled litigation surge" has created a bottleneck in the court system, where the number of judges remains relatively static while the paperwork they must review grows exponentially. Klapper noted that society is at a point where courts are under tremendous strain, with caseloads increasing while resources remain scarce. The cost of litigation is dropping due to AI, which in turn encourages more filings, leaving the judiciary to manage the resulting overflow.
The Los Angeles Superior Court’s pilot program involves a select group of judicial officers who are testing the Learned Hand system across various stages of the litigation lifecycle. This includes everything from the initial intake of a complaint to the finalization of draft rulings. The goal of the pilot is to evaluate the tool’s performance in a real-world environment, ensuring that it provides measurable efficiency gains without introducing errors or bias into the judicial record. Presiding Judge Sergio C. Tapia II emphasized that the court is moving forward with caution, prioritizing the sanctity of the decision-making process. In an official statement, Tapia clarified that while the tool may enhance how judicial officers engage with files, it will not replace or compromise the independence and impartiality of the court. The human judge remains the final arbiter of truth and law.
Technical Safeguards and the Problem of AI Hallucinations
One of the primary concerns surrounding the use of AI in the legal field is the phenomenon of "hallucinations," wherein large language models (LLMs) generate plausible-sounding but entirely fabricated information. The legal world has already seen several high-profile instances of this failure. In 2023, the defense team for Fugees rapper Pras Michel alleged that their own closing argument, which had been partially drafted by AI, included frivolous claims and failed to address key weaknesses in the prosecution’s case. That same year, Michael Cohen, the former attorney for Donald Trump, found himself in legal jeopardy after his lawyers submitted a brief containing citations for court cases that did not exist—a direct result of using an unverified AI tool.
To combat these risks, Learned Hand employs a specialized architecture that prioritizes verification over mere generation. Klapper explained that the system operates within a "walled garden" of source material. Unlike general-purpose AI models like ChatGPT, which draw from the vast and often unreliable expanse of the open internet, Learned Hand is restricted to the specific case filings, evidence, and verified legal databases relevant to a particular matter. This approach, often referred to as Retrieval-Augmented Generation (RAG), ensures that every summary or draft ruling the AI produces is anchored in a specific, verifiable source. Klapper noted that the majority of the expense in their model is dedicated to this verification layer. The system breaks tasks into discrete steps, assigning each step to a model with a specific function to minimize the risk of bias or error.
A Chronology of Judicial AI Development
The adoption of Learned Hand by the Los Angeles Superior Court is the latest milestone in a timeline of technological evolution within the legal sector:
- 2022–2023: The public release of advanced LLMs leads to widespread experimentation by legal professionals. Early adopters face professional sanctions as "hallucinated" citations begin appearing in federal and state court filings.
- 2024: Learned Hand is founded by Shlomo Klapper with the mission of creating "purpose-built" AI for the judiciary, focusing on the specific needs of clerks and judges rather than litigators.
- 2025: Several smaller jurisdictions begin testing AI-assisted administrative tools. The legal community debates the ethics of AI involvement in the "sanctum" of judicial chambers.
- February 2026: The Fisher Phillips report confirms a 49% surge in litigation filings, attributing the increase to the lowered barriers of entry provided by generative AI.
- March 2026: The Los Angeles Superior Court announces its partnership with Learned Hand, launching a formal pilot program to integrate the technology into the nation’s largest trial court system.
Broader Implications for the Future of Justice
The implications of this pilot program extend far beyond the borders of Los Angeles County. If successful, the Learned Hand model could serve as a blueprint for courts worldwide. The primary benefit is the potential for increased access to justice. When courts are backlogged, the time it takes for a case to reach a resolution can span years, often favoring parties with the financial resources to endure long delays. By streamlining the "drudge work" of the judiciary, AI could theoretically allow cases to move through the system more quickly, reducing costs for litigants and the public alike.
However, the integration of AI into the bench also raises profound questions about the future of the legal profession. Traditionally, the tasks now being handled by Learned Hand—summarizing filings and drafting initial memos—have been the training ground for junior attorneys and judicial law clerks. If these roles are increasingly automated, the legal industry may need to rethink how it trains the next generation of jurists. Furthermore, there is the ongoing concern of "automation bias," where human judges might become overly reliant on AI-generated summaries, potentially overlooking subtle but crucial details in the original filings.
To address these concerns, the Los Angeles pilot emphasizes a "point-and-click" interface that requires no technical training or "prompt engineering" from the judges. This design is intended to keep the judge in the driver’s seat, treating the AI as a highly efficient assistant rather than a black-box decision-maker. Klapper’s mantra of "don’t trust, verify" is central to the training provided to participating judicial officers. The AI provides the links and citations for every claim it makes, allowing the judge to verify the accuracy of the output against the original document with a single click.
Conclusion: The Path Toward Augmented Jurisprudence
The Los Angeles Superior Court’s experiment with Learned Hand marks a shift from a reactive stance on AI to a proactive one. Rather than simply banning or restricting the use of the technology, the court is seeking to harness its power to address the very problems the technology itself has helped create. As litigation volumes continue to rise, the traditional methods of judicial administration are reaching their breaking point.
The success of the pilot will likely be measured by its ability to maintain the high standards of the American legal system while meeting the demands of a high-velocity digital age. As Judge Tapia noted, the goal is efficiency and effectiveness, but never at the cost of the "sanctity" of the law. If Learned Hand can prove that it can consistently provide reliable, verifiable support to the bench, it may usher in a new era of augmented jurisprudence—one where the wisdom of the human judge is supported, rather than replaced, by the precision of the machine. The legal world will be watching Los Angeles closely, as the results of this pilot could redefine the role of the judge for the 21st century.
