Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2508.20744

From Law to Gherkin: A Human-Centred Quasi-Experiment on the Quality of LLM-Generated Behavioural Specifications from Food-Safety Regulations

Shabnam Hassani·Mehrdad Sabetzadeh·Daniel Amyot
View original source ↗Read source document ↗Export metadata ↓

Abstract

Context: Laws and regulations increasingly shape software design, development, and quality assurance in regulated domains. Because legal provisions are written in technology-neutral language, deriving concrete specifications, requirements, and acceptance criteria to verify software compliance is difficult and error-prone. Recent advances in generative AI, especially large language models (LLMs), may help automate this process. Objective: We present the first systematic human-subject evaluation of LLMs' ability to derive Gherkin behavioural specifications from legal texts using a quasi-experimental design. Gherkin is a domain-specific language for scenario-based system behaviour descriptions in Given-When-Then form and is well suited to automation in software development. Methods: Ten participants evaluated 60 Gherkin specifications generated from food-safety regulations by Claude and Llama. Each participant assessed 12 specifications across five criteria: relevance, clarity, completeness, singularity, and time savings. Each specification was evaluated by two participants, yielding 120 assessments with quantitative ratings and qualitative feedback. Results: Ratings were uniformly high in the top two categories: relevance 95%, clarity 100%, completeness 94.2%, singularity 93.4%, and time savings 91.7%. No statistically reliable differences were found across participants or between LLMs. Qualitative feedback noted occasional omissions, hallucinations, and mixed intents, underscoring the need for human oversight, especially in safety-critical domains. Conclusion: In food safety, LLMs can assist in deriving Gherkin specifications from legal texts, but omissions and hallucinations require systematic human review.

Explore related subjects

cs.SE

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Shabnam Hassani, Mehrdad Sabetzadeh, Daniel Amyot. 2026-03-11. From Law to Gherkin: A Human-Centred Quasi-Experiment on the Quality of LLM-Generated Behavioural Specifications from Food-Safety Regulations. https://arxiv.org/abs/2508.20744

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

XScientist: A Git-Like Research Protocol for Long-Running Autonomous Scientific Discovery

Autonomous research systems can generate plausible papers while losing the decisions, failed branches, and evidence needed to inspect or continue the work. We present XScientist, a local-first, git-like protocol that treats research state, rather than a manuscript, as the unit of continuation. Hypotheses, experiment attempts, observations, claims, reviews, and handoffs are represented as typed, content-addressed objects in an exploration graph. Immutable checkpoints, explicit negative outcomes, claim-evidence closure, replay boundaries, and authority-aware gates make each transition inspectable without treating a passing integrity check as scientific truth. The protocol exports a portable Agent-Native Research Artifact (ARA) that another agent or human can inspect, fork, verify, and extend. A reference implementation integrates planning, execution, review, repair, and supervised long-running operation while preserving provenance across these stages. We evaluate the protocol with controlled artifact-integrity workloads and matched external task pilots, keeping native task performance separate from evidence and audit claims. The result is an interoperability and accountability layer for long-running autonomous science, with explicit boundaries where human judgment and independent evaluation remain necessary.

cs.SE↗
arXiv2026

AutoSQL: Extracting SQL Templates from Imperative ORM Code in Large-Scale Repositories

Suboptimal SQL queries can significantly degrade the performance of cloud systems, motivating the extraction and auditing of SQL statements before deployment. However, Go ORM frameworks construct SQL imperatively through scattered method-call sequences, making it difficult to statically recover the resulting SQL templates. We present AutoSQL, a system that reconstructs SQL templates from Go ORM code. AutoSQL constructs a Code Index, a directed graph that captures structural dependencies between functions, types, and global variables as navigable edges. It then traces upstream call chains from ORM invocation sites to identify database-interacting functions as entry points. For each entry point, an LLM agent traverses the Code Index to collect code slices that influence SQL generation, switching to pattern-based search when the graph cannot resolve a retrieval goal. We call this strategy Hybrid Context Retrieval. Once sufficient context is collected, the agent synthesizes SQL templates. Evaluation on a benchmark of 579 test-covered entry points and 1,186 runtime-traced SQL statements from five large-scale Go repositories shows that AutoSQL achieves 68.04% to 72.18% recall, exceeding the static reachability baseline by 11.80% to 15.94% and outperforming existing methods by 8.52% to 21.50%.

cs.SE↗
arXiv2026

The Vocabulary of Flaky Tests in Swift

Flaky tests produce non-deterministic outcomes without code change, eroding CI confidence and delaying deliveries. While vocabulary-based machine learning prediction has proven effective for Java and JavaScript, no study has evaluated it for Swift, a language whose testing style is dominated by UI and asynchronous code. We collect 91 flaky and 22,349 stable tests from 15 open-source Swift projects via re-execution and commit-history mining, then train five classifiers (Random Forest, Decision Tree, Naive Bayes, SVM, KNN) on TF-IDF unigram+bigram features under stratified 5-fold cross-validation. Random Forest achieves the best performance (Precision = 0.92, F1 = 0.86, AUC = 0.95) and substantially outperforms trivial baselines, among them a vocabulary-threshold rule applied to the most informative tokens, confirming a genuine discriminative signal (MCC = 0.75 vs. 0.08 for the best baseline). Information-gain analysis reveals two complementary signal types. Flakiness markers appear predominantly in unstable tests and comprise concurrency primitives (async, await), expectation-based synchronisation (expectation, fulfill), error propagation (throws), and explicit timing dependence (timeout, wait, now). Stability markers, chiefly the assertion vocabulary of plainly synchronous tests (xctassertequal), count as evidence against flakiness. Error analysis shows that the model fails when flakiness is hidden in shared infrastructure outside the test body or when async constructs are used in a deterministic context, exposing the intrinsic limit of lexical prediction. These results extend vocabulary-based flakiness detection to the Swift ecosystem and characterise both its effectiveness and its boundaries.

cs.SE↗