Here is the requested information for 'ssrn-4873649' by Professor Yonathan Arbel:

**1. ## TL;DR ≤100 words**

Professor Yonathan Arbel of the University of Alabama School of Law argues that AI's potential to reduce legal costs and increase access to justice paradoxically threatens judicial economy with a litigation boom. Instead of courts historically shrinking rights to cope, he proposes proactively integrating AI tools into the legal system. This would enhance and scale judicial processes, addressing the vast unmet legal needs, leveraging AI's growing capabilities despite current flaws, and preventing regressive responses to increased caseloads. The goal is to improve justice delivery by making the system more efficient and accessible.

**2. ## Section Summaries ≤120 words each**

*   Professor Yonathan Arbel of the University of Alabama School of Law writes that while AI tools offer hope for increased access to justice by sharply reducing the costs of generating legal materials, this very effectiveness paradoxically threatens judicial economy by increasing the volume and verbosity of caseloads. He further writes that rather than courts responding by shrinking substantive rights to manage this influx, as has happened historically, the legal system should proactively integrate AI tools to enhance and scale up the legal process itself.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that a vast number of legal disputes are never filed, with studies suggesting around 120 million legal problems go unresolved in the U.S. each year. He also writes that this access to justice crisis particularly affects low-income Americans, as 92 percent of their significant civil legal issues receive little to no legal aid.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that significant barriers to justice, primarily the high cost of legal services exemplified by average hourly lawyer rates of $292, prevent many individuals from addressing legal problems affecting their basic human needs. He also writes that the sheer investment required means even doubling legal aid budgets has done little to narrow this justice gap, with sociolegal issues like 'legal consciousness' further illustrated by individuals describing being underpaid as being 'stiffed' rather than having their rights violated.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that Nora and David Freeman Engstrom center the access to justice problem on an asymmetry in legal tech adoption, where firms zealously automate litigation while individuals show "anemic adoption" and rely on "analog tools." He also writes that while this argument about tech asymmetry creating power imbalances, particularly in debt collection litigation, has a kernel of truth, the assertion may be too strong or becoming outdated.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that amusing stories of lawyers misusing AI, which support traditional views of the legal profession, distract from the surprising reality that even small firms are adopting these imperfect tools due to their convenience. He also writes that this widespread adoption is anticipated to democratize legal technology, significantly reduce costs, and potentially lead to a litigation boom by expanding access to justice for those currently underserved.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that as AI erodes access barriers, it can precipitate a litigation boom that threatens to overwhelm an already strained judicial system. He further writes that rather than waiting for this crisis, the legal system should proactively integrate AI, despite its current unreliability, to scale up and improve the delivery of justice while carefully considering judicial needs and system constraints.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that early barriers prevent individuals from pursuing legal claims, a process captured by the naming-blaming-claiming model, and AI could help individuals articulate their misfortunes in legally cognizable terms. He also writes that while AI could help access justice, the erosion of these barriers might also lead to a litigation boom of both meritorious and abusive claims, and the essay will use control theory to consider the implications for judicial economy.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that legal doctrines can function as "legal thermostats," where judges adjust procedural and substantive rights to achieve homeostasis, potentially reshuffling rather than solving access to justice problems. He further writes that a proactive solution involves integrating AI tools into the judicial process itself, allowing the legal system to scale up and meet challenges without compromising litigants' substantive rights.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that current-generation AI systems can perform many legal tasks "adequately," which, combined with a large access to justice gap, suggests an impending AI litigation boom. He also writes that AI is neither omnipotent nor a mere trick, and because the technology is rapidly developing, current assessments of its capabilities are merely tentative floors and its limitations are tentative ceilings, with constant improvements ongoing.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that a recent study found AI models like GPT-4 exhibited accuracy in identifying legal issues in contracts on par with junior lawyers, while using only 8% of the time and costing a single quarter compared to a lawyer's average fee of $74.26. He also notes these models, similar to junior lawyers, showed a preference for precision over recall, and their observed mistakes appeared transient and model-specific rather than fundamental, with newer models resolving illustrative errors.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes about a study he co-authored which evaluated large language models as "smart readers" to assist consumers with legal documents like contracts and privacy policies. He further writes that this study found these smart readers significantly reduced contract length and reading time, and improved text readability to a fifth-grade level, without compromising essential information.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that while LLMs show high but inconsistent performance, likely exceeding lay tax preparers, they are poised to substitute for neighborhood representatives like H&R Block or estate planners rather than top-tier lawyers. He also writes that a persistent and "alarmingly prevalent" failure mode is "hallucinations," where LLMs invent non-existent facts, though these can often be checked and advances show promise in reducing this issue.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that while GPT-4 initially appeared to pass the Uniform Bar Exam at the 90th percentile, more refined analyses suggest its performance is closer to the median of test-takers and in the bottom 15th percentile for essays. He also writes that despite the need for caution when extrapolating from exam performance, early real-world studies, such as one where 80% of human referees preferred a GPT-4 drafted complaint letter over one by a trained lawyer, indicate AI's potential effectiveness.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that while current AI models are fast and cheap, rigorous testing shows they perform below the level of median lawyers, and these tests do not fully account for present or future AI advancements like deep prompt engineering. He also writes that the faults found in LLMs must be measured against the realistic alternatives ordinary people have, such as doing nothing, which is a common response, especially for poor households.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that AI demonstrates considerable potential for automating legal tasks, with predictions of automating 44 percent of such work within ten years and early adoption seen in firms like Allen & Overy where a tool called Harvey was quickly used by 25 percent of the practice daily. He also notes that while a 2023 survey found 82 percent of lawyers believe AI can be applied to legal work, only 51 percent deemed it appropriate, and various 2023 surveys reported actual usage among lawyers ranging from 11 to 21 percent.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that while AI adoption in law firms is currently small-to-moderate for tasks like drafting and summarising, it is expected to consistently increase due to developing tools, new lawyers' familiarity, and client pressure for reduced billing. He also writes that the remarkably rapid recent adoption of AI by knowledge workers suggests law firms will not lag for long, indicating AI's utility, potential productivity gains, and a path for broader integration into legal workflows, potentially extending to courts.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that AI offers a holistic shock to the access to justice problem, significantly reducing not only the financial costs of legal services but also addressing social and psychological barriers. He further writes that this impact is crucial because barriers to justice are regressive, and AI can also influence the upstream "naming-blaming-claiming" process through which individual grievances are transformed into recognized legal disputes.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that the 'Naming, Blaming, Claiming' (NBC) filter, which requires individuals to perceive an injury, identify a responsible party, and conceptualize a legal assertion, results in most potential legal claims being lost, with a disproportionately regressive effect on poorer individuals. He also notes that generative AI directly addresses this NBC filter by quickly and cheaply guiding users through all three stages, as illustrated by an AI's response to a tenant's mold query, potentially leading to a radical change in legal consciousness.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that AI models can remove many invisible upstream barriers on the way to justice and assist individuals with legal strategy by making complex information accessible to their specific sociolinguistic needs. He also writes that AI can further aid the litigation journey by drafting required communications and other litigation materials, and help pro se individuals navigate the legal process, thereby giving people more access to justice.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that AI's capacity to reduce access to justice barriers, by aiding in legal research and diminishing litigation uncertainty, strongly suggests an impending "AI litigation boom." He further states that this potential surge in litigation is supported by AI's ability to effortlessly produce verbose legal filings and by strong economic incentives and convenience, which are likely to outweigh initial concerns about the quality of these AI-generated documents.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that AI has the potential to remove barriers to justice, aiding litigants and especially low-income individuals, but its use by strategic players like debt collection firms will likely lead to an AI litigation boom. He also notes that some existing barriers to justice, while filtering important cases, also serve a positive function by deterring vexatious or meritless filings, such as access friction preventing debt buyers from pursuing very small claims.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that the current judicial system, already burdened and demonstrating "lightened scrutiny" under increased caseloads, faces further challenges to its economy with an anticipated AI-driven surge in cases, which historically has led to adjustments affecting primary and procedural rights. He also suggests control theory, which uses feedback mechanisms like a thermostat adjusting temperature, offers a useful framework for understanding how dynamic systems such as the judiciary might regulate themselves to maintain desired states despite disturbances like increased litigation.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that judges regulate the flow of litigation similarly to thermostats, adjusting procedural and substantive doctrines as control mechanisms in response to judicial environment demands like case volume and available resources. He further notes that these common administrative adjustments for judicial economy, such as varying court fees, inevitably affect substantive rights, raising concerns about using legal rights as levers for managing judicial resources.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that mechanisms like fees and de minimis rules, intended to filter low-value cases, can also screen out socially important litigation and disproportionately affect the poor. He also points out that heightened pleading standards, such as those established in *Twombly* and *Iqbal* to control discovery costs, have empirically shown to primarily impact pro se plaintiffs rather than significantly reducing overall filing activity.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that courts employ indirect "procedural thermostats" like Lone Pine orders in complex toxic tort cases, compelling plaintiffs to present preliminary evidence on injury and causation to manage litigation and cull meritless claims, despite criticisms of their burden and inconsistency. He also provides another example of such a litigation control mechanism: the doctrine of exhaustion of administrative remedies, particularly in prisoner's rights cases, which requires plaintiffs to complete agency processes before seeking judicial relief.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that analysis of EEOC data and Lex Machina lawsuit data reveals that only a small percentage of discrimination charges ultimately become federal lawsuits. He further writes that his specific analysis indicates the ratio of unresolved EEOC discrimination claims transforming into actual lawsuits is approximately 3.6 percent.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that legal systems utilize "procedural thermostats," like specific standards of proof and statutes of limitations, which function as regulatory frictions intended to control the volume and quality of litigation. He also writes that these mechanisms operate on the (often unverified) hope that added friction deters less meritorious claims, a premise now challenged by AI tools that can help overcome such procedural barriers.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that AI can ameliorate barriers to justice by assisting people with legal processes, document creation, navigating administrative remedies, and overcoming pleading standards that often filter out cases from less skilled litigants. He also suggests that as AI tools remove such frictions and costs, a commensurate increase in civil litigation, potentially even a doubling in volume due to the large access to justice gap, is not implausible.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that an anticipated AI-driven litigation boom will pressure judicial economy, and while historical responses involved adjusting "legal thermostats," he advocates for proactive integration of AI tools into the judicial process. He also notes that one traditional "legal thermostat" strategy involves increasing fees to reduce filings, but this approach has the significant downside of limiting access to justice for those who cannot afford higher costs.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that despite the rise of litigation financing, fees remain a crude and disproportionately impactful tool for filtering lawsuits, especially for the poor. He also conceptualizes pleading standards as a "proof-of-work" mechanism, akin to blockchain, requiring more upfront effort to filter out weaker claims by leveraging a litigant's own assessment of their case's merits.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that AI writing tools can rapidly transform vague claims into elaborate, well-structured legal arguments, potentially including "hallucinated" facts, making it harder for judges to quickly assess the plausibility of filings, particularly those from pro se litigants. He also explains that this capability undermines "proof-of-work" filtering mechanisms, potentially leading judges to respond by demanding more doctrinally or subtly altering legal standards to conserve judicial resources strained by an increased volume of AI-assisted litigation.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that reactive adjustments to AI's impact, such as increased fees or stricter pleading standards, could paradoxically narrow civil rights and worsen the delivery of justice, as these measures are unstable and not AI-proof. He also notes that alternatively, policymakers might adopt a 'sit and wait' approach, observing AI's development due to technological uncertainty, historical adaptation by the legal system, and unknown AI adoption patterns before making changes.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that despite potential negative uses of AI in law, a passive stance is ill-advised because current trends show increasing AI integration, and its unreliability should serve as a catalyst for careful development and testing. He further argues that this proactive engagement is crucial for refining AI, preparing the judicial system for its benefits, and ensuring it remains at the forefront of technological integration to effectively deliver justice.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that while judicial skepticism towards generative AI is understandable, banning it in courtrooms would be wrong long-term as it would hinder the democratization of access to the justice system and perpetuate existing asymmetries. He also argues that disclosure regimes for AI-generated materials are a hopeless enterprise because reliable detection is unavailable and widespread AI integration will render such disclosures uninformative, akin to disclosing computer use.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that while massively increasing funding is the most direct way to address the growing demand for justice, the realism of a budget increase sufficient to significantly expand judicial capacity appears tenuous in the current political climate. He further notes that even if AI could fully replace legal aid, redirecting its roughly $2.7 billion budget would only afford about a 30 percent increase to the federal court system's $9.4 billion budget, an insufficient sum for a major expansion.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that to effectively manage the anticipated "AI litigation boom," the legal system should pursue an integration strategy, implementing AI in all aspects of the legal process to amplify the productivity of judges and clerks and stretch existing judicial resources. He also notes this integration is already occurring organically, with judges admitting to using AI for tasks like drafting opinions and even for "generative interpretation," indicating such tools are considered useful and will likely become "irresistibly attractive."
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that despite cautious optimism about AI "robo-judging," significant resistance and ethical concerns exist, leading him to argue that focusing on fully automated judging is a distraction. He also states that AI's more powerful utility lies in mundane but impactful applications, such as compressing the immense and increasing volume of text generated in litigation to aid in meting out justice.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that generative AI excels at document summarization, offering two main types: abstractive summarization, which creates new condensed versions conveying core meaning, and extractive summarization, which identifies and compiles key phrases directly from the text. He further explains that both approaches can significantly aid judges, with abstractive summaries directing attention to critical parts and offering robust overviews, while extractive summaries are invaluable for identifying crucial elements like specific evidence or quotes, thereby reducing the material judges need to personally review.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that implementing extractive summarization technologies into case management systems is straightforward and cost-effective, allowing for automated summaries and extraction of key document parts to aid judicial attention management. He also describes a more advanced application, document Q&A, which enables judges to ask specific questions about ingested filings in ordinary language and receive natural language answers from AI based on the document's content.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that document Q&A offers a radical improvement over traditional search by allowing users to ask direct questions in plain language, though it should be viewed as a diligent but fallible assistant. He also notes that while these LLMs may provide partial, misleading, or even hallucinated answers and struggle with complex queries, their limitations can be managed similarly to how judges oversee legal clerks, retaining ultimate responsibility.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that while judges must verify AI-generated facts to ensure judicial efficiency, they can also leverage AI's "generative interpretation" capabilities, which use vast datasets to offer a powerful tool for textualist interpretation surpassing traditional methods. He also notes that confidentiality concerns with cloud-hosted AI models are significant, but evolving solutions like on-premise hosting, data encryption, and the formulation of tailored legal standards are critical for addressing these privacy issues.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that large language models' capacity to discern meaning in context, unlike dictionaries, suggests generative interpretation could be the future of textualist interpretation. He also references Richard Re's analysis which reasonably flags dangers in AI-assisted opinion drafting, such as diminished judicial ownership, deliberation, and authenticity, potentially leading to lifeless boilerplate judgments.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that while judges should be aware of AI's dangers, integrating it carefully into the judicial process, short of outsourcing adjudication, can provide necessary support to meet the potential sharp increase in litigation. He also argues that the real decision is not whether to use algorithms but which kind, as failing to integrate AI could lead to regressive "blind algorithms" like increased fees or stricter standards when judicial resources are strained by AI-driven increased access.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that while AI will significantly reduce barriers to justice, this is merely a prelude to delivering actual justice, and a resulting litigation boom could, like past spikes, lead to a reduction in legal rights. He further proposes the proactive integration of AI tools into the legal process, which, despite concerns about bias that can be managed through careful application, offers a better alternative to bluntly suppressing litigation and calls for lawyers to lead in tool-building scholarship.
*   Professor Yonathan Arbel of the University of Alabama School of Law writes that administrators should spearhead collaboration with technologists to address challenges in the judicial system. He also states that considerations of judicial economy present an urgent choice regarding how much justice society wishes to purchase and whether to extend these resources through automation tools for judges.