AI and the Law: Why a Neural Network Misreads Requirements
Habr AI has published an analysis of why you cannot honestly turn a law into system requirements with a single zero-shot prompt. The model produces a convincing list, but it mixes up roles, drops constraints and keeps no traceability from the legal norm to the design conclusion. Such an answer is useful as a draft and a quick way into the topic, but not as the final result of the analysis.
AI-processed from Habr AI; edited by Hamidun News
A neural network misreads the law: why trying to extract IT system requirements from legislation with a single zero-shot prompt almost always produces false confidence.
Can AI be used to analyze requirements from laws?
You cannot rely on a single zero-shot prompt alone. Between a legal norm and a system requirement lies a whole chain of analytical decisions that a human must make.
Why is zero-shot the wrong approach for analyzing laws?
A regulatory act describes not an interface, a process or an API, but a legal field. A model cannot convert legal norms into system requirements on its own, without analytical work.
Why can't
ChatGPT automatically extract requirements from a law? Because it takes a whole chain of analytical decisions. A law describes a legal field (norms and rules), while a system needs technical requirements (interface, process, API). A neural network cannot carry out this transformation by itself in a single prompt.
Why shouldn't zero-shot be used to analyze laws?
Because between a norm and a system requirement lies a whole chain of analytical decisions that the zero-shot approach does not account for.
Why is a law not a technical specification for a neural network?
A regulatory act describes a legal field: rules and constraints. A technical specification, by contrast, describes an interface, a process or an API. These are different things.
Why does a neural network misread requirements from laws?
Because between a norm and a system requirement lies a whole chain of analytical decisions that cannot be extracted with a single query to an AI.
Is a neural network dangerous for working with legal requirements?
Yes, when people try to extract requirements with a single zero-shot prompt. It creates false confidence that the AI has understood the law correctly.
- Nov 1, 2026Local models lose to Opus and Codex: 0.21 on a MacBook versus 1.00 for Codex
- Nov 1, 2026Kimi scores 96.2% on a 1C database analytics test and emerges as a fallback if Claude access is blocked
- Nov 1, 2026OpenAI Rewrites GPT Behavior Rules: Five Values and a Chain of Command
- Nov 1, 2026Google admits AI Mode cuts search revenue, shares fall 15%
Key Questions Answered
Can AI be used to analyze requirements from laws?
You cannot rely on a single zero-shot prompt alone. Between a legal norm and a system requirement lies a whole chain of analytical decisions that a human must make.
Why is zero-shot the wrong approach for analyzing laws?
A regulatory act describes not an interface, a process or an API, but a legal field. A model cannot convert legal norms into system requirements on its own, without analytical work.
Why can't ChatGPT automatically extract requirements from a law?
Because it takes a whole chain of analytical decisions. A law describes a legal field (norms and rules), while a system needs technical requirements (interface, process, API). A neural network cannot carry out this transformation by itself in a single prompt.
Why shouldn't zero-shot be used to analyze laws?
Because between a norm and a system requirement lies a whole chain of analytical decisions that the zero-shot approach does not account for.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.