ISOprep: an LLM-driven pipeline for semantics-preserving processing of neutralized requirements according to ISO 29148

Kata Amanda Schiller; Arthur Seibel

doi:10.1017/pds.2026.10439

ISOprep: an LLM-driven pipeline for semantics-preserving processing of neutralized requirements according to ISO 29148

Kata Amanda Schiller^*
, Arthur Seibel

^*Korrespondierende/r Autor/-in für diese Arbeit

Professur für Produktentwicklung und Konstruktionslehre

Publikation: Beiträge in Zeitschriften › Konferenzaufsätze in Fachzeitschriften › Forschung › Begutachtung

Abstract

AI in Requirements Engineering (RE) relies on industrial data, yet safety and privacy risks limit its use. While the GDPR distinguishes only between anonymization and pseudonymization, we use neutralization as a semantics-preserving technique. In AI-supported RE, data heterogeneity and cross-domain variability impede model training. We propose guidelines for semantics-preserving preprocessing for RE datasets based on ISO 29148 criteria, showing that neutralization does not compromise semantics. The approach enables industry-academia collaboration through AI-assisted RE in product development.

Originalsprache	Englisch
Zeitschrift	Proceedings of the Design Society
Jahrgang	6
Seiten (von - bis)	807-816
Seitenumfang	10
DOIs	https://doi.org/10.1017/pds.2026.10439
Publikationsstatus	Erschienen - 08.2026
Veranstaltung	19th International Design Conference - DESIGN 2026 - Dubrovnik, Kroatien Dauer: 18.05.2026 → 26.05.2026 Konferenznummer: 19 https://www.designconference.org/

Bibliographische Notiz

Publisher Copyright:
© The Author(s), 2026.

UN SDGs

Dieser Output leistet einen Beitrag zu folgendem(n) Ziel(en) für nachhaltige Entwicklung

SDG 8 – Anständige Arbeitsbedingungen und wirtschaftliches Wachstum
SDG 9 – Industrie, Innovation und Infrastruktur

Fachgebiete und Schlagwörter

Informatik
Ingenieurwissenschaften

ASJC Scopus Sachgebiete

Software.
Modellierung und Simulation
Angewandte Informatik
Computergrafik und computergestütztes Design

Zugriff auf Dokument

10.1017/pds.2026.10439Lizenz: CC BY-NC-ND

Dieses zitieren

@article{7b60ba23794449fe8016c99410c283bf,

title = "ISOprep: an LLM-driven pipeline for semantics-preserving processing of neutralized requirements according to ISO 29148",

abstract = "AI in Requirements Engineering (RE) relies on industrial data, yet safety and privacy risks limit its use. While the GDPR distinguishes only between anonymization and pseudonymization, we use neutralization as a semantics-preserving technique. In AI-supported RE, data heterogeneity and cross-domain variability impede model training. We propose guidelines for semantics-preserving preprocessing for RE datasets based on ISO 29148 criteria, showing that neutralization does not compromise semantics. The approach enables industry-academia collaboration through AI-assisted RE in product development.",

keywords = "Informatics, requirements management, artificial intelligence (AI), neutralisation, ISO 29148, data‑driven design, Engineering",

author = "Schiller, \{Kata Amanda\} and Arthur Seibel",

note = "Publisher Copyright: {\textcopyright} The Author(s), 2026.; 19th International Design Conference - DESIGN 2026, DESIGN 2026 ; Conference date: 18-05-2026 Through 26-05-2026",

year = "2026",

month = aug,

doi = "10.1017/pds.2026.10439",

language = "English",

volume = "6",

pages = "807--816",

journal = "Proceedings of the Design Society",

issn = "2732-527X",

publisher = "Cambridge University Press",

url = "https://www.designconference.org/",

}

TY - JOUR

T1 - ISOprep

T2 - 19th International Design Conference - DESIGN 2026

AU - Schiller, Kata Amanda

AU - Seibel, Arthur

N1 - Conference code: 19

PY - 2026/8

Y1 - 2026/8

N2 - AI in Requirements Engineering (RE) relies on industrial data, yet safety and privacy risks limit its use. While the GDPR distinguishes only between anonymization and pseudonymization, we use neutralization as a semantics-preserving technique. In AI-supported RE, data heterogeneity and cross-domain variability impede model training. We propose guidelines for semantics-preserving preprocessing for RE datasets based on ISO 29148 criteria, showing that neutralization does not compromise semantics. The approach enables industry-academia collaboration through AI-assisted RE in product development.

AB - AI in Requirements Engineering (RE) relies on industrial data, yet safety and privacy risks limit its use. While the GDPR distinguishes only between anonymization and pseudonymization, we use neutralization as a semantics-preserving technique. In AI-supported RE, data heterogeneity and cross-domain variability impede model training. We propose guidelines for semantics-preserving preprocessing for RE datasets based on ISO 29148 criteria, showing that neutralization does not compromise semantics. The approach enables industry-academia collaboration through AI-assisted RE in product development.

KW - Informatics

KW - requirements management

KW - artificial intelligence (AI)

KW - neutralisation

KW - ISO 29148

KW - data‑driven design

KW - Engineering

UR - https://www.scopus.com/pages/publications/105045271731

U2 - 10.1017/pds.2026.10439

DO - 10.1017/pds.2026.10439

M3 - Conference article in journal

AN - SCOPUS:105045271731

SN - 2732-527X

VL - 6

SP - 807

EP - 816

JO - Proceedings of the Design Society

JF - Proceedings of the Design Society

Y2 - 18 May 2026 through 26 May 2026

ER -

ISOprep: an LLM-driven pipeline for semantics-preserving processing of neutralized requirements according to ISO 29148

Abstract

Bibliographische Notiz

UN SDGs

Fachgebiete und Schlagwörter

ASJC Scopus Sachgebiete

Zugriff auf Dokument

Fingerprint

Dieses zitieren