From telegraphs to GPT-4

AI: The new tool in the legal toolbox

2024 February

Legal scholar Roscoe Pound warned that “the law must be stable, but it must not stand still.” The same is true for the practice of law. The journey from the quill’s scratch to the silent hum of quantum computation has been marked by a cycle of reluctance, acceptance, and dependency.

Legal tools and technologies in the 19th and 20th centuries

In the 19th century’s industrial bloom, the electric telegraph stretched wires like nerves, transmitting legal thought faster than any horse-bound courier. Lawyers, guardians of tradition, eyed this electric interloper warily until its undeniable expediency won them over. The typewriter’s clack soon followed, casting aside the flowing ink of legal manuscripts. It clashed with the artisanal craft of law but ultimately reigned, creating a new canon of legal documentation – impersonal, uniform, efficient.

Amplifying commerce in the Roaring Twenties, the telephone cut through the legal profession’s reservations, its rings heralding a new era of immediacy in counsel and consultation. This tool, once an oddity, became as central as the gavel in orchestrating the rhythms of justice. Photocopying, too, made a stealthy entrance amid the mid-century’s machine age, facing initial distrust. However, its rapid-fire replication of legal texts eventually underpinned an era where case law could be duplicated as swiftly as it was cited.

The 1970s bore witness to the personal computer’s incursion into the legal sanctum, a leap from ledger to logic board. The profession’s old guard cast a skeptical eye on this digital dossier, yet the allure of streamlined data proved irresistible. The subsequent digital revolution, spearheaded by Westlaw and LexisNexis, redefined research. These databases, numb to the touch of paper and tome, offered a legal compendium unfettered by geography or time. The laborious and time-intensive exercise of legal research was now outsourced to the algorithm, with reliance on the practitioner’s memory passing to the machine’s infallible index.

Artificial Intelligence (AI) and the law

As of the end of 2023, the AI landscape boasts the likes of GPT-4 and its digital kin – each offering a suite of capabilities, but also brimming with potential pitfalls. No currently available AI, despite its sophistication, is a panacea for legal judgment. Misinterpretation, hallucinations, confidentiality concerns, ethical blind spots, and the erosion of the practitioner’s skill stand as ready vices promising to sabotage those who surrender their discernment to the seduction of technology.

GPT-4’s role in law today is that of an ancillary – a digital aide-de-camp for the attorney. It serves as a scribe for preliminary research, a drafter of base documents, an editor for legal prose, and a statistical oracle for discerning data-driven insights. It is a tool that complements but does not replace the nuanced expertise of the seasoned attorney, when used appropriately.

The advent of AI, particularly GPT-4, has opened new avenues in personal injury law. Attorneys now have access to tools that can efficiently analyze medical records, accident reports, and other critical case documents. This capability not only saves time but also uncovers insights that might be missed by the human eye.

In assessing damages and liabilities, AI can provide more accurate estimations based on vast datasets of past cases. This helps in formulating more effective case strategies, negotiations, and in advising clients more accurately about their chances of success and potential compensation.

While AI presents exciting possibilities, it also raises ethical questions, including in personal injury law. The use of AI must align with the principles of justice and fairness. Concerns about data privacy, particularly with sensitive medical records, and the potential biases in AI systems must be addressed. California attorneys must be vigilant in ensuring that their use of AI complies with legal standards and ethical norms.

Like its disruptive predecessors, AI will be met with reluctance and skepticism by certain segments of the legal industry due to the pitfalls that can stem from its employment. However, absent societal collapse, the ubiquitous integration of AI into the legal profession is inevitable. Those who fail to learn how to use it, how to avoid its vices, how to utilize its virtues, and to keep up with improvements to the software will fall far behind.

AI in personal injury

In personal-injury cases, GPT-4’s utility is not just theoretical, but increasingly practical. These tools have evolved to analyze extensive PDF document troves, synthesizing their contents into concise summaries and targeted responses. Consider, for example, their application in dissecting deposition transcripts, incident reports, DME reports, and medical records. They distill these often voluminous materials into digestible formats, providing quick access to key information. This capability, while still maturing, is advancing beyond mere foundational stages, showing promising strides toward enhancing legal efficiency.

In the intricate and often labor- intensive process of written discovery, the application of AI technology marks a significant advancement in efficiency. Utilizing tools like GPT-4, attorneys can navigate the vast seas of discovery documents with unprecedented speed and precision. This technology excels in extracting pertinent information from the depths of extensive data sets – a task traditionally demanding considerable time and human resources. By efficiently identifying critical evidence, discerning patterns in communications, and even offering insights into the probable tactics of opposing counsel, AI enhances our ability to focus on the substantive aspects of legal strategy. Sophisticated use of this new technology can expedite repetitive tasks seen in discovery requests, responses, and meet-and-confer letters.

The integration of AI in this phase of discovery should not diminish the attorney’s role, but rather, refine and quicken it, enabling us to allocate our skills and judgment more effectively. In essence, the incorporation of AI into written discovery aligns with a forward- thinking legal practice, where technology serves as a complement to, rather than a replacement for, the nuanced and critical thinking that is the hallmark of effective lawyering.

Beyond mere document analysis, AI’s predictive capabilities offer valuable insights into case outcomes. By analyzing historical data, these tools aid attorneys in setting realistic client expectations and crafting more informed case strategies. However, it’s crucial to underscore that AI, as it stands, cannot replicate the nuanced grasp of human suffering, empathy, or moral judgment – elements at the heart of personal-injury law. Herein lies an essential role of the attorney. Yet, AI can support this human element by aiding in articulating these nuanced aspects. It can assist in crafting compelling narratives, identifying resonant analogies, and shaping themes that underpin persuasive demand letters, cogent closing arguments, and effective deposition preparations.

Fundamentally, a lawyer is trained to apply facts to law – a skill honed from law school onwards. This quintessential aspect of legal analysis, often structured in the IRAC format (Issue, Rule, Application, Conclusion), is amenable to and enhanced by AI technologies. AI can assist in structuring legal arguments, identifying pertinent legal principles, and applying facts to these principles in a coherent and logical manner, thereby reinforcing the foundational skills of legal analysis.

Basic description of the current technology

ChatGPT, developed by OpenAI, is based on the Generative Pre-trained Transformer (GPT) architecture. It is the most advanced broad-purpose AI language model available to the public, and known for its ability to generate human-like text. ChatGPT is trained on a diverse range of internet text sources, allowing it to generate responses across various topics. Its design includes multiple layers of neural networks, which help in understanding and generating text based on the input it receives.

The model is trained through a process called unsupervised learning, where it learns to predict the next word in a sentence without explicit guidance. This extensive training enables ChatGPT to understand context, answer questions, and create content that can mimic human writing styles. ChatGPT processes input text, comprehends the context, and generates relevant text-based responses. It’s capable of handling a broad spectrum of language-related tasks, including conversation, answering queries, and text generation.

For those hesitant to engage GPT-4, it’s important to recognize its advancements over GPT-3. GPT-4 surpasses its predecessor in scale and complexity, enabling more nuanced text processing and generation. Its extensive training on a diverse range of internet text sources enhances its accuracy across various topics. Notably, GPT-4 shows improved understanding of context in conversations, leading to more coherent and relevant responses.

While efforts to reduce biases and inaccuracies – concerns with GPT-3 – have progressed in GPT-4, challenges remain. Additionally, GPT-4 introduces multimodal capabilities, a step beyond text generation, and offers enhanced customization options for diverse professional applications. In essence, GPT-4 represents a significant evolution in natural language processing, expanding its potential applications and effectiveness in complex linguistic tasks.

ChatGPT’s strengths lie in its versatile text generation, capable of creating diverse content like essays, stories, and code, and its adeptness at understanding multiple languages and styles due to extensive training. It’s particularly effective in real-time interactions, offering prompt responses, which proves useful in dynamic communication scenarios.

However, ChatGPT’s limitations are notable. It lacks real-world understanding, relying solely on its training data and therefore missing current context and personal experiences. Users should be wary of inherent biases in its responses, a reflection of its internet-based training data. Additionally, while ChatGPT handles a range of tasks, it struggles with complex topics that demand expert knowledge or deep reasoning. The quality of its responses is also directly tied to the clarity and structure of the input it receives, with vague or poorly formulated queries often leading to subpar responses.

Challenges of implementing AI

The integration of AI into personal injury law is not without its challenges. Ignorant or blind reliance on the technology is a malpractice case waiting to happen. Levidow, Levidow & Oberman, P.C. has become emblematic of this problem when, in March of 2023, the attorneys for that firm “abandoned their responsibilities when they submitted non-existent judicial opinions with fake quotes and citations created by the artificial intelligence tool ChatGPT” according to U.S. District Judge Kevin Castel in an order issuing sanctions of $5,000.

Such instances reveal that AI, though sophisticated, still lacks the nuanced judgment and reliability of competent legal counsel. The technology frequently produces ‘hallucinations’ – confident assertions of facts and sources that simply do not exist. As lawyers, the onus falls on us to discern the chaff from the wheat, to critically evaluate the output of these tools. Mastery over AI in legal practice involves not just using the technology, but understanding its mechanics and limitations.

There is a temptation among practitioners to await the arrival of user-friendly AI tools – a passive approach that may prove shortsighted. While such tools will no doubt democratize the use of AI in law, they might also commoditize certain legal services. Today’s legal practitioners must consider how to integrate AI into their practice innovatively, to maintain and enhance the value of their services.

Furthermore, it would be misguided to expect that legislative or regulatory bodies will impede the integration of AI in law, particularly when such technologies promise to bridge the access-to-justice gap. As personal injury lawyers, we must reckon with the potential impact of AI on traditional practice models, including contingency-fee structures. While it is true that some areas of legal practice (i.e., contracts, wills and trusts) will be more susceptible to earlier replacement than others (i.e., trial advocacy), it does seem apparent that the way things are done today will not be capable of competing with the way things will be done tomorrow.

Change invites opportunities

These changes simply invite new opportunities to adjust to new vehicles of legal practice. Imagine a NASCAR driver, accustomed to the high speeds and oval tracks of their sport, is presented with the opportunity to pilot a Formula 1 car. The NASCAR driver, despite their expertise and experience, confronts a distinct set of challenges in an F1 car. Formula 1 racing, with its advanced technological intricacies, different aerodynamic principles, and distinct strategic approach, necessitates a specialized skill set and a deep understanding of the vehicle’s capabilities and handling. The NASCAR driver must adapt their skills, learn the nuances of the F1 car’s performance, and refine their approach to meet the demands of this new racing discipline. This transition is not merely a step up; it’s a step into a different realm, requiring dedicated practice, an open mind to continuous learning, and the flexibility to adjust to a novel racing environment.

In the legal sphere, the integration of AI, particularly for an attorney well-versed in traditional methods, presents a similar challenge. AI in law, offering robust tools for data analysis, case prediction, and research, requires more than just a foundational understanding of legal principles. It demands a specialized knowledge of how these technologies function, an appreciation of their appropriate application in various legal contexts, and a keen awareness of their ethical boundaries. The attorney must not only embrace the potential of AI to augment their practice, but also remain vigilant of its limitations and pitfalls. It is a journey of adaptation, where continuous education and skill refinement are paramount, and it is necessary to evolve with the times while steadfastly adhering to the core values of our profession.

Indeed, ethical considerations are paramount. The integration of AI in law necessitates a vigilant adherence to professional standards, including confidentiality, competent legal analysis, and compliance with legal and ethical norms. As set forth by the California Rules of Professional Conduct, lawyers have a duty to remain abreast of changes in law and its practice, including technological advancements. Rule 1.1 describes the duty of competence, and the comments to that rule explicitly state, “The duties set forth in this rule include the duty to keep abreast of the changes in the law and its practice, including the benefits and risks associated with relevant technology.” (Emphasis supplied).

Despite these challenges, the potential benefits of AI in legal practice are significant. AI can enhance case management, refine legal research, and boost overall efficiency, allowing lawyers to focus on the strategic elements of litigation. The true potential of AI in consumer advocacy lies not merely in technical expertise, but in linguistic agility and familiarity with legal process – both areas where lawyers, as masters of language and legal nuance, can truly excel.

Yet, this is not to say that the path is devoid of obstacles. The lawyer’s traditional role is under renegotiation in this new digital age. To stay relevant and effective, the legal practitioner must become a hybrid, blending the time-honored proficiency in the art of persuasion with the astuteness of a technologist. This synthesis of old-world advocacy and new-age acumen might just be the linchpin in the evolution of legal practice. Here, the younger lawyer may have an advantage over her seniors.

In the dance of this sequence, GPT-4 and its ilk represent not a challenger, but a partner to the legal choreography, a digital consort to the eloquent pas de deux of litigation. As the profession steps into this rhythm, it must do so with eyes wide open, acknowledging the technology’s limitations while embracing its possibilities. The prudent lawyer will thus don the dual mantle of advocate and innovator, leveraging artificial intelligence to elevate the practice of law, ensuring that justice, not just efficiency, is served.

Conclusion

In this era of transformation, the prudent will discern the call to adapt. The legal vanguard will not be those who merely adopt new tools but those who integrate them into the very sinew of their practice, crafting a new jurisprudence that is as forward-looking as it is deeply rooted in the bedrock principles of the law.