By Daniel Pelc, Director of Client Solutions and Integration, HaystackID
I come from a sprawling family of physicians and physicists. Years of dinner-table osmosis left me reasonably medically literate, or so I thought.
A recent CT scan put that to the test. The radiologist’s read came back: “Bones: No acute or suspicious osseous abnormality. Severe facet arthropathy at L4-L5 and L5-S1. Thoracic DISH.” The first sentence read clean. The second one shook me. Severe = severe-pathy = disease. I was quietly planning my imminent demise.
I entered the report into a generative AI (GenAI) engine. It told me, in plain English, that nothing is alarming in my lower vertebrae and that I have substantial arthritis in my lower back. The panic the radiology report induced. The calm the GenAI translation produced. That gap between clinical language and human understanding is the same one healthcare-adjacent counsel now confronts in every malpractice and internal-investigation matter.
Traditionally, that gap has been bridged through expert witnesses, consultants, and painstaking document review. GenAI is changing that. But it is also introducing new obligations, new risks, and a question the legal profession has not yet fully answered: What does competent representation look like when the underlying record is clinical, and the translator is a machine?
A fool for a client, revisited
The phrase “a person who represents themselves has a fool for a client” has been passed down through over 17 generations of attorneys since William De Britaine first printed it in 1682. It endures because the underlying logic is sound: expertise has limits, and recognizing those limits is itself a form of competence. The same logic now applies to the relationship between law and medicine and increasingly, to the tools attorneys use to bridge the two.
Almost as entrenched is the belief, held by nearly every attorney except seemingly those in intellectual property and a few other outliers, that law and science do not mix. Judge William Schwarzer of the U.S. District Court for the Northern District of California put it plainly: science and technology issues in litigation “challenge the ability of judges and juries to comprehend the issues — and the evidence — and to deal with them in informed and effective ways. As a result, they tend to complicate the litigation, increase expense and delay, and jeopardize the quality of judicial and jury decision making.”
That was written before GenAI existed. The gap Judge Schwarzer described has not closed, but the tools available to work across it have changed considerably.
What Rule 1.1 requires now
Rule 1.1 of the ABA Model Rules of Professional Conduct requires that a lawyer provide competent representation, defined as “the legal knowledge, skill, thoroughness and preparation reasonably necessary for the representation.” In 2012, the ABA revised Comment 8 to make explicit that competence also requires lawyers to “keep abreast of changes in the law and its practice, including the benefits and risks associated with relevant technology.” Most U.S. jurisdictions have since adopted that language or issued analogous guidance. ABA Formal Opinion 512, issued July 29, 2024, extended that framework directly to GenAI, addressing competence, confidentiality, client communication, candor toward the tribunal, supervisory duties, and reasonable fees.
The trajectory is clear. Technological fluency is no longer optional for legal practice. For counsel handling clinical records, malpractice defense, and healthcare investigations, it is the competence standard.
The Rosetta Stone problem
GenAI is starting to operate as a kind of Rosetta Stone for cryptic medical minutiae. It can rapidly translate dense radiology reports, operative notes, and discharge summaries into something a non-physician can interrogate. But the value proposition goes beyond simplification. Hidden in that minutiae may be exculpatory evidence that shores up a defense or confirms a deviation from the standard of care in the case of an internal investigation.
In medical malpractice litigation and healthcare-related internal investigations, the practical applications are accelerating. AI-assisted medical record review tools can ingest record sets exceeding 1,000 pages, build litigation-grade chronologies linked to source pages, and surface inconsistencies between provider notes. In years past, the distillation of these concepts into actionable information that once consumed dozens of associate hours can now be performed in minutes. GenAI can correlate clinical findings with published literature, stress-test causation theories, and cluster similar incidents across departments to surface patterns that manual review would miss.
The Thomson Reuters Institute’s 2025 Generative AI in Professional Services report, based on 1,702 respondents, found that 26% of legal organizations are actively using GenAI, up from 14% in 2024, with document review, legal research, and document summarization as the top three use cases. Seventy-eight percent of law firm respondents said they expect GenAI to become central to their workflow within five years.
Fast is not the same as right
Augmentation, not replacement, is the only defensible posture. The attorneys most likely to handle this transition well are not the ones who hand a GenAI engine a question and blindly accept the answer. Successful practitioners treat GenAI output as a starting hypothesis, then verify it against primary records, treating physician interviews, and credentialed expert review. Build the verification step into the workflow before finalizing case hypotheses and before the deadline pressure arrives. Document which queries were run, which model was used, which outputs were corroborated against source material, and which were discarded. Federal judges speaking at the Thirteenth Biennial Bench-Bar Conference on April 24, 2026, in Greenbelt, MD, have been candid:
“For judges, the problem is especially serious because court opinions and orders carry legal authority. A mistaken citation in a judicial opinion can confuse litigants, mislead attorneys, and create unnecessary appellate issues. For lawyers, the risk is just as high: attorneys who submit AI-generated work without verifying every citation may face sanctions, reputational harm, or professional discipline. AI can help organize research, summarize documents, and identify issues, but it cannot be treated as a substitute for legal judgment. Every authority cited in a brief or opinion must still be checked against reliable legal databases.”
The deskilling concern runs in both directions. Medical educators have warned that clinicians who lean on GenAI translation may lose the diagnostic instinct that comes from reading the raw report. The risk of GenAI hallucinations affecting medical diagnostics should send a chill through any physician. The same risk applies to attorneys who stop interrogating the underlying record.
The risk is already in the courtroom
GenAI can hallucinate, oversimplify, or miss nuance in highly specialized clinical contexts. In litigation or investigation settings, those errors carry material consequences. In Mata v. Avianca, decided in June 2023, U.S. District Judge P. Kevin Castel sanctioned two New York attorneys with $5,000 for submitting a brief containing fictitious case citations generated by ChatGPT, opinions the attorneys never read, because the cases did not exist. In 2026 alone, there have been 979 court decisions relating to AI-generated case law citations. Sanctions have escalated from warnings to monetary fines, mandatory training, bar referrals, and exclusion from representation.
GenAI outputs must be treated as starting points, not conclusions. Human expertise, both legal and medical, remains the final arbiter.
The compliance questions that GenAI fails to answer
The adoption curve carries a privacy and information-governance dimension that healthcare-adjacent legal teams cannot ignore. Patient records routed through a GenAI tool are subject to HIPAA, and any vendor processing protected health information on behalf of a covered entity must have a Business Associate Agreement in place before a single record is uploaded. Vendor contracts should explicitly bar the use of client data for training public models.
Whether AI-assisted summaries and chronologies are work product, discoverable, or both remains an open question that depends on how the output was generated, by whom, and for what purpose. Judges have been candid that the answer will turn on the workflow documentation, not the tool’s marketing copy.
Rule 1.1’s mandate of thoroughness and preparation, read alongside Comment 8 and Formal Opinion 512, now requires technological fluency not as a luxury but as a duty. The better modern aphorism may be this: a lawyer who refuses to use available tools may not be a fool, but they may be increasingly outmatched.
We would never accept a surgeon acting as their own malpractice lawyer mid-procedure. The stakes are too high, the expertise too specialized. Yet the inverse expectation has been quietly accepted for years: that lawyers can navigate dense medical issues without truly speaking the language. GenAI does not eliminate that gap. It narrows it in a meaningful way, enabling better questions, sharper analysis, and more informed advocacy.
Medicine will only become more complex. Data will only become more voluminous. The expectation that lawyers can independently and competently interpret that complexity without technological assistance is becoming unrealistic. Where does your practice draw the line between AI-assisted competence and AI-dependent malpractice, and who in your organization is responsible for keeping that line current?
News sources
- Quote Origin: A Man Who Is His Own Lawyer Has a Fool for a Client (Quote Investigator)
- Patent Law and the Two Cultures (120 Yale L.J. 2, 2010) (Yale Law Journal)
- Rule 1.1 Competence — Comment (American Bar Association)
- ABA Formal Opinion 512: Generative Artificial Intelligence Tools (July 29, 2024) (American Bar Association)
- Mata v. Avianca, Inc., No. 1:22-cv-01461, Dkt. No. 54 (S.D.N.Y. 2023) (Justia)
- AI Hallucinations in Court: Every Lawyer Needs to Read This Before Their Next Filing, Legal AI World, March 2026
- Federal Judges Split on AI in Courts as Use Grows and Errors Mount, JD Journal, April 27, 2026
- AI Hallucinations in Court Filings and Orders: A 2025 Review of Sanctions Across the Courts and Rule Proposals (Sterne Kessler)
- Thomson Reuters Survey: Over 95% of Legal Professionals Expect Gen AI to Become Central to Workflow Within Five Years (LawSites)
- Remembering Judge William W. Schwarzer (1925-2017) (U.S. District Court Northern District of California Historical Society)
About HaystackID®
HaystackID® solves complex data challenges related to legal, compliance, regulatory, and cyber requirements. Core offerings include Global Advisory, Cybersecurity, Core Intelligence AI™, and ReviewRight® Global Managed Review, supported by its unified CoreFlex™ service interface and eDiscovery AI™ technology. Recognized globally by industry leaders, including Chambers, Gartner, IDC, and Legaltech News, HaystackID helps corporations and legal practices manage data gravity, where information demands action, and workflow gravity, where critical requirements demand coordinated expertise, delivering innovative solutions with a continual focus on security, privacy, and integrity. Learn more at HaystackID.com.
Assisted by GAI and LLM technologies.
SOURCE: HaystackID
Advisory Note: As legal matters increasingly involve complex medical, clinical, and healthcare data, technology competency is becoming inseparable from legal competency. Generative AI offers new opportunities to accelerate record review, enhance issue identification, and improve case preparation, but its value depends on disciplined workflows, expert oversight, and defensible processes. Organizations operating at the intersection of law, healthcare, and compliance must balance innovation with accuracy, privacy, and professional responsibility. HaystackID helps clients operationalize AI-enabled workflows through secure data management, advanced analytics, expert-led review, and governance frameworks that support informed decision-making while maintaining the standards required by courts, regulators, and clients.
