The Disadvantages of Artificial Intelligence in the Legal Profession: What Lawyers Need to Know
Artificial intelligence is no longer a distant concept in the legal profession — it is already reshaping how law firms operate, how legal research gets done, and how legal documents are drafted, reviewed, and delivered to clients. Platforms powered by generative AI, including widely used tools like ChatGPT, Claude & Perplexity, have made it easier than ever for attorneys to process information at scale, automate repetitive tasks, and streamline workflows that once took hours.
But as with every wave of new technologies to sweep through the legal industry, the adoption of AI systems requires far more than enthusiasm. It demands rigor, accountability, and a clear-eyed understanding of the limitations of AI — because in the legal field, the consequences of getting it wrong can be severe.
At NBI, we have spent over four decades helping legal professionals stay ahead of the curve through continuing legal education. And increasingly, that means helping attorneys understand not just the benefits of AI, but its very real disadvantages. Here is what every lawyer, paralegal, and law firm administrator needs to understand before leaning too heavily on AI tools in their practice.
AI Outputs Are Not Always Accurate — and the Errors Can Be Catastrophic
Perhaps the most widely documented disadvantage of AI in the legal field is the problem of hallucination. Large language models — the AI models that power tools like ChatGPT and many legal AI platforms — are trained on vast datasets and generate responses by predicting likely text sequences. They do not "look up" facts the way a human researching a case would. As a result, AI-generated content can include fabricated citations, invented case names, and legal conclusions that sound authoritative but are entirely wrong.
This is not a theoretical risk. Attorneys have already faced sanctions for submitting AI-generated briefs containing nonexistent case citations that were not verified before filing. The responsibility for the accuracy of any document submitted to a court — or provided to a client as legal advice — rests with the attorney, not the algorithm. No matter how capable an AI tool appears, human oversight is not optional. It is a professional and ethical obligation.
For legal research specifically, this creates a significant challenge. Legal research has always required precision: the right citation, the correct holding, the accurate procedural history. AI systems that produce confident-sounding but unreliable outputs are potentially more dangerous in legal research than a gap in knowledge, because they may not signal their own uncertainty.
The Ethical Considerations Are Complex and Evolving
The use of AI in legal practice raises profound ethical considerations that the legal profession is still working through. The American Bar Association (ABA) and numerous state bar associations are actively developing guidance, but the rules of professional conduct were written long before generative AI tools existed. Attorneys are left to interpret existing obligations — competence, confidentiality, supervision, and candor — in a context those rules did not anticipate.
Competence, under Model Rule 1.1, requires lawyers to understand the technology they use. The ABA has made clear that this includes AI technology. But what does competent use of a large language model actually look like in practice? How should attorneys evaluate AI outputs, check citations, and assess whether a summary of a contract or statute is accurate? These are questions that demand real answers — not assumptions.
The bar association landscape is equally fragmented. New York, California, Florida, and other jurisdictions have issued varying levels of guidance on AI use in legal practice. Some state bars have formed dedicated AI task forces. Others have been slower to respond to advancements in generative AI tools. This patchwork of rules creates confusion for attorneys practicing across jurisdictions and underscores why ongoing education on AI in law is essential.
Client Confidentiality and Data Privacy Are Serious Concerns
When attorneys input client information into an AI system, where does that data go? This question sits at the heart of one of the most pressing disadvantages of AI in the legal industry today.
Many generative AI tools, including consumer-facing chatbots, are designed to learn from user interactions. Entering client data — factual summaries, case details, identifying information — into these platforms may expose that information in ways that violate attorney-client privilege and run afoul of an attorney's duty to protect client confidentiality. Even enterprise-grade legal AI platforms require careful review of data handling practices, terms of service, and security certifications before client information is entrusted to them.
Data privacy obligations extend beyond ethical rules. Attorneys handling legal issues involving healthcare, financial services, or family law deal with sensitive client information that is subject to state and federal privacy regulations. The use of AI tools that process, store, or transmit this data without adequate safeguards creates potential legal exposure not just for the client, but for the firm. Law firms must conduct due diligence before deploying any AI-powered solution that touches client data — and in many cases, that scrutiny needs to be even more rigorous than what is applied to other third-party service providers.
AI Cannot Replace Human Judgment or Contextual Understanding
Even the most sophisticated AI models lack something that every experienced attorney develops over years of practice: contextual understanding. The law is not a lookup table. Legal work requires interpreting ambiguity, weighing competing interests, applying judgment to novel facts, and understanding how a particular judge, jurisdiction, or opposing counsel is likely to respond. These are functions that AI systems are not capable of performing reliably.
Document review offers a clear example. AI-powered tools can dramatically accelerate the process of sorting through thousands of documents in discovery or contract analysis. But they can miss context that only a trained attorney would catch: the significance of an unusual clause, the strategic implications of a particular admission, the relationship between a document and a client's broader legal exposure. Automation in document review should be understood as a filter, not a replacement for legal judgment.
The same applies to decision-making at every stage of legal practice. AI tools can generate options, surface patterns, and produce summaries. But the attorney must evaluate those outputs, apply professional judgment, and take responsibility for the conclusions reached. Delegating decision-making to an algorithm — even a sophisticated one — is not consistent with the professional obligations lawyers carry.
The Risk of Bias in AI Systems
The algorithms that power legal AI tools are trained on historical datasets, and those datasets reflect the world as it has been — including its inequities. AI systems used in legal contexts have the potential to replicate and even amplify biases embedded in the data they were trained on. This is a documented concern in areas ranging from criminal justice to contract analysis to predictive legal research.
For law firms committed to equitable representation and ethical practice, the potential risks of bias in AI outputs deserve serious attention. Neither the legal profession nor the legal industry at large has fully reckoned with how machine learning models can embed discriminatory patterns — and how those patterns can influence real-world legal work without anyone recognizing what is happening.
Supervision, Competence, and the Duty to Understand What You Are Using
Thomson Reuters, Westlaw, LexisNexis, and other established legal research platforms have introduced AI-powered features into their products. Meanwhile, newer legal AI companies are entering the market with bold claims about their AI capabilities. In this environment, it is easy for attorneys and law firms to adopt AI tools without fully understanding how those tools work, what their limitations are, or how to evaluate the quality of their outputs.
The ABA's guidance makes clear that competence extends to supervising staff — including paralegals — who use AI tools in the course of their work. Attorneys cannot disclaim responsibility for AI-assisted work product simply because they did not personally operate the tool. Supervision of AI use within a firm requires the same diligence as supervision of any other legal work.
This has significant implications for law firm leadership and training initiatives. Firms that adopt AI tools without investing in attorney education — on both the capabilities and the limitations of AI — expose themselves to malpractice risk, disciplinary exposure, and the erosion of client trust.
What the Disadvantages of AI Mean for Your Practice
None of this means attorneys should avoid AI entirely. The benefits of AI — greater efficiency in legal research, faster document review, the ability to streamline time-consuming functions — are real and increasingly difficult to ignore. The legal profession is not going to opt out of this technological shift. But responsible adoption requires understanding that AI is a tool, not a practitioner.
The limitations of AI in legal practice are not bugs to be patched in the next update. They are inherent to how these systems work. Generative AI tools, including the most advanced AI models currently available, do not understand the law. They generate text. The attorney who uses that text bears full professional responsibility for what it says.
Staying ahead of these challenges means investing in your own understanding of AI — its technical foundations, its ethical implications, and its practical limitations. It means reviewing your firm's policies on AI use, client confidentiality, and data privacy. And it means treating the ethical concerns raised by AI in law not as future problems to address someday, but as current obligations that require attention now.
Earn CLE Credit While Navigating AI in Your Practice
NBI offers a growing library of CLE courses covering AI in law — from foundational AI skills and ethics to advanced topics in legal research, workflow integration, and the rules of professional conduct as they apply to AI use. Whether you are in New York or navigating requirements in another jurisdiction, our accredited courses are designed to give you practical, actionable knowledge you can apply immediately.
The legal field is changing. The attorneys who thrive will be those who engage with that change thoughtfully, critically, and with a clear understanding of both the promise and the peril of AI in legal practice.