Published On: July 10, 2026
Category: Choosing a lawyer
Artificial intelligence has made legal information more accessible than ever. With a few prompts, someone can ask an AI tool whether they have viable lawsuit, have a contract drafted, summarize a court decision, or research a potential legal claim. But accessible information is not the same as reliable legal advice.
Increasingly, prospective clients arrive at our office having done their own research on the merits of their potential litigation. Their research has been created and organized by AI. Existing clients now attempt to advise us on how to manage our cases or draft documents, claiming that “AI has all the answers.” The unfortunate reality is that AI is causing nightmares for our prospective and current clients because it is often filled with consequential flaws, incorrect advice, and is based on biased or inadequate prompts, incomplete context, or information that may not reflect the current law in the relevant jurisdiction.
In this article, we will discuss the current state of AI-generated legal conclusions and the concerns that exist for both non-attorneys and attorneys. AI is constantly evolving, and attorneys will need to adapt their practices as these tools become more capable. But for people attempting to handle legal matters themselves, the fundamental problem remains: AI can provide information about the law without offering the judgment, verification, and strategic analysis required to apply that law to a specific case.
Sorry to say to most of the DIY folks out there, AI is not going to replace attorneys anytime soon.[1]
Can AI Provide Reliable Legal Advice?
AI can be useful for learning about legal concepts, generating questions for an attorney, organizing information, or helping someone understand an unfamiliar document. The problem begins when a person treats an AI-generated response as a legal conclusion rather than as a starting point for further research.
A legal answer depends on more than the question presented. It can depend on the jurisdiction, the specific facts, the procedural posture of the case, the applicable statute, controlling precedent, contractual language, and facts the person asking the question may not realize are legally significant.
Two people can ask an AI tool essentially the same question and receive different answers because small differences in the facts or the way the prompt is framed can change the analysis. That is one reason AI-generated legal research should not be treated as a substitute for advice from a qualified attorney
AI is Like Magic
Arthur C. Clarke said that “Any sufficiently advanced technology is indistinguishable from magic.”[2] When I first started playing around with AI a couple of years ago, it felt a lot like a psychic had picked me out of the audience to read my mind. I learned quickly, however, that AI had its limits, especially when it came to anything related to the law, and the news of attorneys getting in trouble around the country for using it was becoming scary.[3]
While AI tools appear sophisticated, they lack the verification capabilities, strategic judgment, and current legal knowledge that professional legal representation provides. It may be a good starting point for considering and understanding legal concepts, but it will not replace professional legal judgment, verification, and strategy. I am not saying that just because I am an attorney. So, let’s see why this is.
What Legal Information Does AI Actually Have Access To?
The important question for someone seeking legal advice is not simply whether an AI system has been trained on a large amount of information. It is whether the system has access to the specific, current, and authoritative legal sources needed to answer the question being asked.
AI is dependent primarily on (1) public internet content, rather than private or premium legal databases (e.g., Westlaw, Lexis) and (2) free legal resources, including government websites and published case law. While this is a massive amount of data, there are limitations due to time constraints (meaning the case law may not be up to date), gaps in quality control, and potential bias.[4]
Hallucinations and Fabricated Legal Authorities
Consequently, the use of AI can result in “hallucination or fabrication.”[5] Additionally, AI systems can sometimes produce citations to cases, statutes, or other legal authorities that do not support the proposition for which they are cited, or in some circumstances, authorities that do not exist at all.
That risk is particularly serious in legal research. A fabricated citation may look convincing to someone without legal training, especially when it is presented alongside a confident explanation of the law.
AI is also not inherently aware of jurisdictional boundaries. A response may discuss California law when the relevant case is governed by Colorado law, cite federal law when state law controls, or rely on an outdated rule without adequately identifying the problem.
The Problem with Prompts
The origins of AI emerged in the 1950s and, until recently, had sputtered and faltered, being mainly the product of science fiction rather than fact. A significant challenge has been integrating artificial intelligence with computational limitations and natural language.[6]
Garbage In, Garbage Out
Prompts, or user inputs, are crucial to the accuracy of AI. The prompt is critical because AI systems generate responses based on the information and instructions provided to them; they do not independently know which facts have been omitted or which assumptions are incorrect.
What does that mean? I have many clients come in with their AI analysis of their case. Many prompts look like this: “Can I sue for this?” This prompt, standing alone, will always come with the answer: “yes.” An AI-generated answer can only be as useful as the information and assumptions underlying the question. If legally significant facts are missing, the response may be confidently wrong without the user realizing why
Another example is when someone asks an AI to draft a “simple” contract between parties in two different countries using a translation program. These contracts are especially problematic due to nuanced language differences, jurisdictional differences, and other elements that are not addressed in the contract. The result is a defective or inaccurate contract that appears adequate to an inexperienced client but is flawed by the limitations of AI.
Context is Missing
DIY legal research cannot interact with an attorney who can assess credibility, clarify information, and dig deeper for intentional or unintentional disregard of damaging information or exaggeration about the situation. As most attorneys know, the case your client presents to you at the beginning is their audition to “win” you as their attorney.
AI is unlikely to discern the missing pieces, and in situations involving trials before a jury or a court, it will not be able to assess how your client will fare in their case on a personal level.
It also cannot independently interview witnesses, evaluate their credibility, inspect physical evidence, observe a client’s demeanor, negotiate with opposing counsel, or make the hundreds of judgment calls that arise during litigation.
Framing
How the prompt is framed will change the output (and often raises ethical issues at the same time). For instance, “How can I get out of this contract?” is a very different question from “What provisions in the contract address termination or enforceability of the contract?”
The second question is more neutral and focused, but even a well-framed prompt does not eliminate the need to verify the resulting legal analysis. The quality of the question matters, but so does the quality and completeness of the information used to answer it.
Careful framing of the question, with tools that are law-specific (Westlaw, Nexis), and the assistance of an experienced attorney will enable you to more adequately (and ethically) guide you in a contract dispute.
The Attorney Difference
A licensed attorney provides value that AI cannot replicate, including, but not limited to: applying legal rules to specific facts, identifying risks and weaknesses, and developing a strategy tailored to the client’s goals.
An attorney also takes professional responsibility for the advice provided and can adjust that advice as the facts, law, and circumstances change
Prompt Usage
An attorney’s legal training, background, and experience allow for more accurate and complete prompts through an understanding of:
- The information that is legally and factually relevant to the client’s case.
- How to frame questions to get helpful strategic guidance.
- When to ask follow-up questions to clarify AI responses.
- How to verify and cross-check AI-generated information.
- Which facts may undermine the client’s preferred interpretation of the case.
Verification
Attorneys know how to verify legal information through reliable databases (including Westlaw and/or Lexis, state databases, proprietary sites, and other resources), understand the hierarchy of legal authority, and can distinguish between binding and persuasive precedent.
An attorney does not simply ask whether an AI-generated answer sounds reasonable. The attorney must determine whether the legal authority actually supports the conclusion and whether that authority applies to the client’s jurisdiction and circumstances.
Strategic Thinking – “Know When to Hold Them, Know When to Fold Them.”
Legal training develops judgment about when to fight, when to negotiate, and when to walk away – decisions that require understanding consequences AI cannot foresee.
A legally viable claim is not necessarily a claim worth pursuing. Litigation involves costs, risks, timing, business relationships, evidence, potential recovery, and the likelihood of success. An attorney can evaluate those competing considerations when determining how to proceed.
Ethical Obligations
Attorneys are bound by professional responsibility rules and malpractice insurance, providing accountability and recourse that AI platforms cannot offer.
Attorneys also have professional obligations regarding competence, confidentiality, conflicts of interest, and the handling of client information. Those obligations are fundamentally different from relying on a consumer-facing AI platform for legal guidance.
Relationship Management
Legal matters often involve ongoing relationships, institutional knowledge, and negotiation dynamics that require human insight and experience.
An attorney can communicate directly with the client, opposing counsel, witnesses, experts, courts, and other participants in a legal matter. That human interaction can reveal information and strategic considerations that no standalone AI response can fully capture.
Can AI Replace a Lawyer?
AI is likely to change how attorneys research, draft, organize, and manage legal work. It may make certain tasks faster and allow attorneys to devote more time to higher-level analysis and client service. But using AI to perform a task is different from transferring responsibility for a legal matter to an AI system.
The distinction matters because clients do not simply need information about the law. They need someone who can determine which law applies, evaluate the facts, identify risks, develop a strategy, and take responsibility for executing that strategy.
The Risks of Relying on AI for Legal Advice
AI can be a starting point for understanding legal concepts, but it cannot and should not replace professional legal judgment, verification, and strategy. The law is complex, nuanced, and constantly evolving in ways that require human expertise to navigate safely.
Use AI Legal Information as a Starting Point, Not a Final Answer
AI can be useful for preparing questions, learning basic terminology, or organizing information before speaking with an attorney. It becomes much more dangerous when a person relies on an AI-generated answer to make a decision affecting a lawsuit, contract, business, financial interest, or other significant legal right.
Before making decisions that could affect your legal rights, financial interests, or personal freedom, consult with a qualified attorney who can provide verified, jurisdiction-specific, and strategically sound legal guidance.
The cost of proper legal consultation is usually much lower than the cost of fixing mistakes made by relying on AI for legal research. As a litigator, the author can assure you that having your contract, transaction, or other legal analysis done at the outset of your endeavors will be far less than litigation, which can run in the hundreds of thousands of dollars, years of sleepless nights, and adverse outcomes. Your legal matters deserve the protection that only qualified professional representation can provide.
Contact Lantz Law for Legal Guidance
When the stakes are high, the question is not whether AI can give you an answer. The question is whether that answer is accurate, current, applicable to your jurisdiction, and appropriate for the facts of your situation. Those are questions that require professional legal judgment.
If you are facing a legal issue or are unsure about whether AI-generated information applies to your situation, contact Lantz Law for experienced, jurisdiction-specific legal guidance before making decisions that could affect your rights or interests.
OTHER RESOURCES
In addition to the materials referenced in the article, the author suggests these books for further reading on AI and ethics:
Paedra Boinordiris, Beth Ruddin, Peter Scott, AI for the Rest of Us (2023).
Annette Buhler, Navigating Ethical Leadership in the Age of AI: The Ethic Pocketknife (2024).
DISCLAIMER
The information contained in this document and available on the website lantzlawgroup.com is for general informational and educational purposes only. The information does not contain legal advice. Reading the information in this document or any materials on the lantzlawgroup.com website does not create an attorney-client relationship between you and the author(s) or Lantz Law Group. This document is not to be used as a substitute for competent legal counsel from a licensed provisional attorney in your state or country. If you need legal assistance regarding any of the content contained in this document, please get in touch with a competent and licensed attorney for help.
[1] Ken Crutchfield, AI Won’t Replace Lawyers – But it Will Change How they Work, Above the Law, https://abovethelaw.com/2024/03/ai-wont-replace-lawyers-but-it-will-change-how-they-work/ (March 26, 2025).
[2] Arthur C. Clarke was a renowned English science fiction writer, futurist, and inventor, best known for his novel 2001: A Space Odyssey and his visionary ideas about space travel.
[3] Debra Cassens Weiss, Confronted with AI Hallucinations in Filings, One Court Shows ‘Justifiable Kindness,’ While Another Gets Tough, ABA Journal https://techstory.in/lawyer-fined-5500-sent-to-ai-school-for-fabricated-chatgpt-caselaw/#google_vignette (August 19, 2025).
[4] Cait Levin, AI Prompt Engineering: How to Instruct AI Models to Avoid AI Bias and Get the Results You Want, https://techstory.in/lawyer-fined-5500-sent-to-ai-school-for-fabricated-chatgpt-caselaw/#google_vignette (September 3, 2024).
[5] What are AI Hallucinations, IBM, https://www.ibm.com/think/topics/ai-hallucinations (last visited August 25, 2025).
[6] Tery Winograd and Fernando Flores, Understanding Computers and Cognition – A New Foundation for Design (1st Ed. 1986, 1987)