AI Can Process Law. It Can’t Practice It.

Advances in generative artificial intelligence (“GAI”) have demonstrated a genuine capacity to identify legal issues, generate plausible arguments, summarize doctrine, synthesize precedent, and predict litigation outcomes with increasing sophistication. For many observers, GAI’s rapid evolution suggests that legal expertise itself may be approaching automation.

The prediction is understandable. It’s also built on a faulty foundation.

To believe that lawyers will be replaced by machine automation assumes that law is primarily an act of information processing, when in fact it’s an exercise of contemplation and accountability.

Law Exists Where Rules Are Insufficient

If there were a world in which legal rules determined their own application with perfect clarity one hundred percent of the time, lawyers would be largely unnecessary.

The existence of courts, advocacy, judicial interpretation, appellate review, and procedural discretion is key evidence that this is not the case. These structures reflect a critical reality about law: rules do not apply themselves. Instead, facts are incomplete. Circumstances differ. Principles conflict. Human consequences matter. Legal institutions exist specifically because societies require mechanisms for interpreting ambiguity in conditions where certainty is unavailable. This is not a flaw waiting to be solved by carefully crafted prompts or more sophisticated technology. It is the nature of law itself.

Law operates at the intersection of the structure of rules and the complexities of human existence. It attempts to create coherence across circumstances that cannot be fully anticipated in advance. Legal questions become most significant precisely at the point where the mechanical application of “this is what you do when…” ends and where the discernment of “who,” “what,” “when,” “where,” “why,” and "how" becomes necessary.

That’s why practicing law has never just been about knowing the law.

What lawyers actually do is interpret institutional systems on behalf of people whose lives will be shaped by how those systems operate. They help individuals and organizations navigate environments where the “technically available” option, the “strategically advisable” option, and the “humanly sustainable” option are not always the same thing.

This requires more than data retrieval. It requires judgment about consequences, legitimacy, risk, responsibility, and meaning within human contexts.

Information Is Not the Same Thing as Understanding

Generative AI systems can model legal language with remarkable sophistication. What they can’t do is understand what legal participation means for the people subject to institutional authority. This distinction matters because legal questions are not purely computational.

They involve legitimacy—whether people experience institutional decisions as fair, coherent, and trustworthy.

They involve accountability—who bears responsibility when decisions produce harm.

They involve interpretation—how individuals understand the consequences of the choices available to them.

And they involve judgment—how competing obligations should be weighed when rules alone do not produce a self-evident answer.

These are not simply information problems. They’re human governance problems.

A system may predict how a court is likely to rule with increasing accuracy while remaining incapable of determining whether the ruling meaningfully resolves the underlying conflict, protects legitimate interests, or aligns with broader principles of justice and institutional trust. 

Prediction is not discernment.

Output is not responsibility.

And information is not understanding.

The Most Important Work Lawyers Perform Is Often Outside of Legal Doctrine

Clients rarely arrive with neatly categorized legal issues. They arrive with situations—uncertainty about what is happening, fear about consequences, competing obligations, damaged relationships, institutional confusion, financial pressures, reputational risk, and/or deeply personal stakes tied to identity, family, livelihood, and stability.

The work of legal practice begins with interpretation, long before doctrine is applied.

Lawyers help clients determine what is actually at issue, which facts matter, what risks are real, what tradeoffs exist, what can realistically happen, and which options align with their realities and contexts. This is not merely technical analysis. It’s structured sense-making.

Now of course AI can do some of this, but what it cannot do is bear responsibility for helping another human being decide which consequences they are prepared to live with. This difference is not sentimental. It’s structural.

It requires translating legal complexity into questions and guidance that empowers people to make informed decisions with clarity rather than panic or confusion. It requires understanding that legal outcomes do not occur in isolation and instead occur within interconnected systems of families, workplaces, communities, organizations, and institutions that continue to exist after the matter itself concludes.

The Crucial Question Is Not One of Capability

Much of the conversation surrounding AI and the legal profession has been framed as a debate about capability: Can AI do legal work?

Increasingly, the answer is yes—at least with respect to many “lower level” legal functions, such as research, citations, document drafting—but this framing ignores the more important question: Is practicing law simply about performing legal tasks?

These are not the same thing. 

Performing legal tasks can be automated and has been long before GAI became a talking point (think back to the invention of spellcheck, email, printers, the Internet, and other technological advances). Practicing law, on the other hand, involves responsibility-bearing interpretation within systems of authority on behalf of people who must live with the consequences of institutional decisions.

That responsibility cannot be automated.

Someone must ultimately evaluate the reliability of information, determine which considerations matter most, assess consequences beyond procedural outcomes, exercise discretion where rules are incomplete, and remain accountable when those judgments prove flawed.

This is why recent AI sanctions cases are more significant than they initially appear. [1] The central issue in each of these cases is not just FRCP Rule 11(b) violations for hallucinated case citations—it’s about delegated judgment and owning one’s actions. [2]

These courts—and those to follow—are effectively reaffirming a foundational principle of legal practice: the licensed attorney remains responsible for the reasoning, representations, and authority submitted under their name, and the outcomes, regardless of which tool assisted in generating them. None of that changes just because the tool happens to be one that gathers, processes, and generates information in seconds. 

AI systems do not bear consequences for errors. They cannot be sanctioned, disbarred, held fiscally accountable, or ethically responsible for institutional harm. Lawyers can, which means that the core function of legal practice remains human judgment exercised under conditions of accountability.[3]

What AI Means for the Future of the Profession

There is no doubt that AI will continue to transform the economics and structure of legal work. It’s almost guaranteed that the use of GAI specifically will be guided and constrained through legislation, precedent, and industry standards. Regardless of what guardrails are implemented, the success of such guardrails—their application, review, adjudication, and evolution—will still require human discernment and accountability, bringing us right back where we started.

Rather than eliminate the need for lawyers, the shifting scale and potential of GAI clarifies what legal expertise actually is. It’s forcing legal institutions to distinguish between the work that is fundamentally mechanical and the work that has always required human judgment, ethics, and understanding. It’s exposing the difference between simply processing information and bearing responsibility for how that information is used.

The lawyers most equipped for the future will not be those attempting to compete with large-scale machine processing. It will be those capable of exercising judgment where systems, rules, and technologies are insufficient on their own, because law has never been solely about answers. It’s always been about interpretation, legitimacy, accountability, and the responsible exercise of authority in circumstances where certainty is unavailable and consequences are human.

AI hasn’t changed that. It simply made it clearer.

[1] See Wilder v. President & Fellows of Harvard Coll., Nos. 2384CV01461-BLS2, 2384CV01389-BLS2, 2026 Mass. Super. LEXIS 30, at *7 (Mass. Super. Ct. May 18, 2026) (barring an attorney from appearing pro hac vice in Mass. after he was sanctioned in Wyo. for filing legal papers containing AI-hallucinated case citations, with the court emphasizing that his failure to demonstrate meaningful safeguards against future violations was “surprising [a]nd troubling.”); see also Miller v. Regions Bank, No. 2:24-cv-1324-HDM, 2026 U.S. Dist. LEXIS 114487 (N.D. Ala. May 21, 2026) (highlighting that the profession's critical role in distilling and analyzing case law is “not solely rooted in some high-minded ideal of the legal profession as an honorable one,” but rather a specific duty to clients and the community at large).

[2] Miller, 2026 U.S. Dist. LEXIS 114487, at *36 (distinguishing that “[l]awyers make errors. Competent and ethical lawyers own them”).

[3] See Davis v. Marion Cty. Superior Court Juv. Det. Ctr., No. 1:24-cv-01918-JRO-MJD, 2026 U.S. Dist. LEXIS 111607, at *12 (S.D. Ind. May 20, 2026) (asserting that the legal process is a “semi-sacred conversation” between individuals who have a “personal interest” in the proceedings and their outcomes).

Diana Morris works at the intersection of governance, decision-making, and human judgment. A Chief of Staff and law student with a background in higher education leadership, compliance, and executive advisory work, she examines how individuals and institutions interpret rules, exercise authority, and make sound decisions in complex environments.