evidoria

← Back to browse

Good practice Imported

ICT-Berufsbildung Schweiz Uses Custom GPT Models to Draft National VET Curricula and Exam Questions

Switzerland · Bern · See the Switzerland profile · See the Bern profile

Evidence: Descriptive / self-reported Top 89% 27/100 · Ask Evidence Copilot about this practice

Switzerland's national ICT vocational body has used custom GPT models since 2024 to draft competence-based curricula and generate exam questions, cross-checked by an AI exam lab — though an OECD 2026 survey finds only one in five Swiss VET providers report using AI at all.

20 (7 of 35 respondents) %
Swiss VET providers reporting any AI use in curriculum/qualification development (OECD 2026 survey)
50 %
Providers citing satisfaction with current procedures as a reason for non-use (OECD 2026 survey)
46 %
Providers citing limited capabilities to use AI as a reason for non-use (OECD 2026 survey)
29 %
Providers citing security, privacy or data-ownership concerns (OECD 2026 survey)
ICT-Berufsbildung Schweiz Uses Custom GPT Models to Draft National VET Curricula and Exam Questions

Details

Promoter
ICT-Berufsbildung Schweiz
Period
2024–2025
Keywords
vocational education, ICT curriculum development, professional qualifications

Context

ICT-Berufsbildung Schweiz is Switzerland's national professional organisation for ICT vocational education, leading the development of competence-based, modular curricula and federal assessments delivered in the country's three official languages. Its modules are published on the public Modulbaukasten platform, used by training providers and industry experts nationwide to design curricula and prepare candidates for federal ICT examinations.

Activities

Since 2024, the organisation has used customised GPT models to help industry expert groups draft and document competence-based frameworks defining the required skills, knowledge and performance outcomes for each ICT profession. AI tools are also used to generate and update open-ended exam questions and scenarios; an AI-powered exam lab cross-checks this AI-generated content for quality control before it reaches candidates. The organisation's own site confirms it is actively building out AI-related training offerings, including a new applied-AI qualification, alongside this operational use.

Results

An OECD (2026) national survey found this remains an outlier rather than the norm: of 35 Swiss VET provider respondents, only 7 (20%) reported using AI in curriculum or qualification development at all, in line with the ten-country average. The most common reasons providers gave for not using AI were being satisfied with current procedures (50%) and limited capabilities to use AI (46%), with security, privacy and data-ownership concerns cited by 29%. Government-wide restrictions on using public AI systems with non-public data further limit experimentation, particularly within ministries; OECD interviews found no system-wide AI use for qualification development as of September 2025, with adoption varying widely by canton and institution.

Conclusions

ICT-Berufsbildung Schweiz's own operational use of customised GPT models for curriculum and exam-content drafting is a documented outlier within the Swiss VET sector, not the sector norm.

Implementation
Indicative cost
Low (< €50k)
Time to results
Medium (1–3 years)
Staffing & skills
ICT-Berufsbildung Schweiz (national ICT VET body), industry expert groups drafting competence frameworks, AI-powered exam lab (quality-control function)

Conditions for success

  • existing modular, competence-based curriculum infrastructure (Modulbaukasten) to build AI assistance into
  • trilingual national delivery system already in place
  • a built-in quality-control step (the exam lab) cross-checking AI-generated content before it reaches candidates

Common failure modes

  • OECD's national survey finds the broader Swiss VET sector reluctant to adopt AI in curriculum development (only 20% of providers), citing satisfaction with current procedures (50%), limited capability (46%) and security/privacy/data-ownership concerns (29%)
  • government-wide restrictions on using public AI systems with non-public data limit wider experimentation, especially within ministries
  • no system-wide AI use for qualification development was found as of September 2025, with adoption varying widely by canton and institution

Where it fits

Governance type
national industry/professional VET body
Scale
single national sector body; broader VET ecosystem lagging
Income level
high-income country (Switzerland)

Commonly funded by

National / regional programmes Own resources / municipal budget

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Replication kit

Reusable artefacts from this practice — as published by their sources.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful