Law After the Prompt
Why legal education may change less—and more—than expected
The illusion of automation
Few domains seem as suited to artificial intelligence as law. It is built on text, structured through rules and accumulated over decades of precedent. Legal systems produce vast quantities of documents—case law, statutes, opinions, contracts—all of which can be processed, analysed and recombined. From the outside, law appears almost algorithmic. And increasingly, it behaves that way.
Artificial intelligence can now scan jurisprudence in seconds, identify patterns across thousands of cases and generate structured legal arguments with remarkable fluency. What once required hours of research and drafting can now be produced almost instantly.
This creates a powerful impression. That legal reasoning itself is becoming automatable. But that impression is misleading.
“Legal adjudication is not an algorithmic exercise, but a guarantee of fundamental rights.”
Advocates General, Court of Justice of the European Union
Law is not simply the application of rules to facts. It is the interpretation of norms within context, under conditions of uncertainty and with consequences that extend beyond the case itself.
What appears as logic is, in reality, judgment.
The rise of the machine paralegal
Artificial intelligence is exceptionally effective at what might be called the preparatory layer of legal work.
It can identify relevant case law, compare legal arguments, detect inconsistencies and generate draft documents that are structurally sound and linguistically precise. In many respects, it functions as the ultimate paralegal—faster, tireless and increasingly sophisticated.
This is not a marginal shift. It is a structural one.
Legal professionals are already integrating these capabilities into their workflows, not as experimental tools but as standard instruments of practice.
“AI will not replace lawyers, but lawyers who use AI will replace those who do not.”
Richard Susskind, Author of Tomorrow’s Lawyers
The implication is clear. The profession is changing. But the question is how far that change reaches.
Where AI stops
Artificial intelligence can analyse, generate and suggest. It can organise complexity and simulate reasoning. But it does not decide.
It does not weigh competing values. It does not interpret ambiguity in light of societal context. It does not carry responsibility for the outcome. This is not a limitation that can simply be engineered away. It is structural.
“Judicial decisions require human responsibility and must remain subject to human oversight.”
European Commission for the Efficiency of Justice (CEPEJ), Council of Europe
Law is not only about what can be justified within a system of rules. It is about what should be justified and why. That distinction cannot be reduced to data.
The difference between correct and just
This is where the core tension becomes visible. Artificial intelligence can produce legally coherent answers. It can generate arguments that align with precedent, follow established reasoning and appear internally consistent. But coherence is not the same as justice.
A legally correct answer is not necessarily a just decision. Because law does not operate in a vacuum. It operates within a normative framework that includes proportionality, fairness and the protection of fundamental rights.
These are not variables. They are judgments. And judgment requires accountability.
Simulated reasoning
This creates a new kind of risk. Not that AI produces obviously incorrect outcomes, but that it produces outcomes that are plausibly correct. Convincing. Structured. Difficult to challenge on the surface.
The danger is not that AI is wrong. It is that it can be credibly wrong. Legal reasoning can be simulated without being grounded in responsibility. Arguments can be constructed without anyone fully owning them.
This introduces a subtle but profound shift. From reasoning as a human act to reasoning as a generated artefact. And that shift matters. Because in law, the legitimacy of a decision depends not only on its structure, but on the authority behind it.
The crisis of casuistry
Legal expertise has always been developed through exposure to cases. Through repetition, variation and gradual refinement, students and young professionals learn how to interpret, argue and decide. They develop what might be called legal intuition—the ability to recognise what matters in a specific situation.
This process depends on engagement with the “messy middle” of legal work.
The ambiguous cases.
The imperfect arguments.
The slow construction of reasoning.
If artificial intelligence increasingly absorbs this layer—handling the research, structuring the argument and producing the first draft—then a critical question emerges: Where do future lawyers learn the craft?
If the path to expertise is smoothed out, the formation of judgment becomes less visible and potentially less robust.
Implications for legal education
This is where the impact becomes most visible. Legal education has traditionally placed strong emphasis on knowledge: knowing case law, understanding doctrine, applying rules to hypothetical scenarios.
But if access to legal information is no longer a constraint, that emphasis becomes insufficient.
“The role of legal education is not to produce legal information, but to train legal reasoning.”
Faculty of Law, University of Oxford
The focus must shift. Not away from knowledge, but beyond it.
Towards interpretation.
Towards argumentation.
Towards the ability to justify decisions under conditions of uncertainty.
Students must learn not only how to construct arguments, but how to take ownership of them.
Less change, more pressure
Paradoxically, legal education may change less than expected—and more than expected at the same time.
Less, because the core of legal reasoning remains human. The need to interpret, to weigh and to decide does not disappear.
More, because everything around that core is accelerating. Research becomes faster. Drafting becomes easier. Access becomes universal.
This increases the pressure on what remains. The human element is no longer supported by slower processes. It is exposed.
Europe’s legal moment
In this context, Europe occupies a distinctive position. Its legal systems are deeply rooted in principles of rights, proportionality and institutional accountability. The emergence of AI does not occur in a vacuum, but within a framework that already emphasises limits. This is reflected in the European approach to regulation.
“High-risk AI systems shall be designed and developed in such a way that they can be effectively overseen by natural persons.”
European Union Artificial Intelligence Act, Article 14
Human oversight is not presented as an option. It is a requirement. This reinforces a fundamental idea: that responsibility cannot be delegated to systems, regardless of their capability.
Redesigning legal education
If the role of the lawyer shifts from producing information to exercising judgment, legal education must adapt accordingly.
This does not mean abandoning doctrine or analysis. It means rebalancing the curriculum. Less emphasis on reproducing knowledge. More emphasis on:
- defending arguments in real time
- engaging with ethical dilemmas
- working through ambiguity
- understanding the limits of automated reasoning
Students must be trained not only to arrive at answers, but to justify them—to explain why a particular interpretation holds and what consequences it entails.
Law after AI
Artificial intelligence will continue to transform legal work. It will make processes more efficient, analysis more comprehensive and access more widespread. But it will not change what law ultimately is.
Law is not a system for producing answers. It is a system for making decisions under norms. And those decisions carry consequences that cannot be outsourced.
Final line
Artificial intelligence may assist legal reasoning, but it cannot carry legal responsibility.
This article is part of the series The University After AI, published in the Culture & Education section of Altair Media.
Photo by Sebastian Pichler on Unsplash
A court built on columns of law and judgment—reminding us that while systems may assist reasoning, responsibility remains human.
