AI Use of Covered Code and Documents
Appendix to CET Developer License Agreement
Last Updated: September 23, 2026
Drafting intent: permit ordinary developer AI assistance, including prompting and retrieval-augmented generation, while preserving CSAB confidentiality, IP and trade secret protection and prohibiting unauthorized model training.
| Use Case | Position | Condition |
| Prompting / AI Assistance | Permitted | Allowed through Approved AI Services, provided no Model Training occurs. |
| RAG and retrieval indexes | Permitted | Allowed for retrieval only, including vector indexes and embeddings used solely for retrieval. |
| Private AI Model Training | Requires separate approval | Requires CSAB prior written approval and a separate written agreement. |
| Public AI Model Training | Prohibited | Never permitted, directly or indirectly, including open-weight or third-party model improvement. |
This Appendix forms part of the CET Developer License Agreement ("CETDL") between Configura Sverige AB ("CSAB") and Licensee.
Unless expressly modified herein, all provisions of the CETDL remain in full force and effect.
In the event of a conflict between this Appendix and Sections 2.7 or 6 of the CETDL, this Appendix shall govern solely with respect to the permitted AI uses expressly described herein.
1. Purpose
The purpose of this Appendix is to permit certain artificial intelligence assisted development activities involving Covered Code and Documents while preserving CSAB's intellectual property rights, trade secrets, confidential information and proprietary interests.
All rights not expressly granted in this Appendix remain reserved by CSAB.
2. Definitions
2.1 Covered Materials
"Covered Materials" means all Covered Code, Original Code, Modifications, Source Code, Documents and other materials protected under the CETDL.
2.2 AI Assistance
"AI Assistance" means use of an artificial intelligence system to search, analyze, explain, summarize, troubleshoot, document, review, test, refactor, comment upon, or otherwise assist development activities with respect to Covered Materials, provided that such use does not constitute Model Training.
2.3 Retrieval-Augmented Generation
"Retrieval-Augmented Generation" or "RAG" means a system whereby Covered Materials are indexed, searched, retrieved and temporarily provided to an AI model as contextual information for responding to user queries, without updating or modifying the underlying model weights or parameters.
For avoidance of doubt, RAG may include search indexes, vector indexes, embedding stores, retrieval databases and similar retrieval mechanisms, provided such mechanisms are used solely for retrieval and are not used for Model Training.
2.4 Model Training
"Model Training" means any activity that causes an artificial intelligence or machine learning model to learn from, adapt to, optimize against, incorporate, or otherwise modify its weights, parameters or behavior using Covered Materials.
Model Training includes pre-training, supervised learning, unsupervised learning, reinforcement learning, transfer learning, fine-tuning, instruction tuning, adapter training, embedding model training, synthetic dataset generation for training purposes, reinforcement learning from human feedback and any comparable process resulting in permanent modification of model behavior.
2.5 Approved AI Service
"Approved AI Service" means an artificial intelligence service, platform, model, application or API that has been approved by CSAB and listed in the then-current Approved AI Services Schedule attached to this Appendix or otherwise communicated by CSAB in writing.
To qualify as an Approved AI Service, the service must, at a minimum:
- contractually protect Covered Materials as confidential information;
- not use Covered Materials to train, fine-tune or improve Public AI Models;
- maintain commercially reasonable security safeguards;
- provide access controls appropriate for enterprise use;
- permit use consistent with this Appendix; and
- satisfy any additional requirements established by CSAB from time to time
- Covered Materials remain confidential;
- Covered Materials are not used to improve Public AI Models;
- access is restricted to authorized personnel; and
- outputs and model artifacts remain under Licensee's control.
- Covered Materials may be used for training, optimization or improvement;
- model weights may be distributed outside Licensee's organization;
- Covered Materials or Covered-Material-derived information may become available to third parties; or
- CSAB cannot reasonably verify that Covered Materials are excluded from future model training activities.
- CSAB's prior written approval; and
- a separate written agreement executed by CSAB and Licensee governing such Model Training.
- train a Public AI Model;
- fine-tune a Public AI Model;
- improve a Public AI Model;
- contribute to any public training dataset;
- create synthetic training datasets for Public AI Models;
- provide Covered Materials for use in public model optimization; or
- otherwise cause Covered Materials or Covered-Material-derived knowledge to become incorporated into a Public AI Model.
2.6 Private AI Environment
"Private AI Environment" means an AI environment operated solely for Licensee's internal use in which:
2.7 Public AI Model
"Public AI Model" means any artificial intelligence model, service, platform or system for which:
2.8 Approved AI Services Schedule
The Approved AI Services Schedule may be updated by CSAB from time to time.
The addition, removal or modification of Approved AI Services shall not constitute an amendment to this Agreement.
Licensee is responsible for ensuring that any AI service used with Covered Materials remains listed as an Approved AI Service at the time of use.
3. Permitted AI Uses
Subject to all other terms of the CETDL, Licensee may use Covered Materials for AI Assistance.
Permitted AI Assistance includes code explanation, debugging, troubleshooting, code review, test generation, documentation generation, architectural analysis, code navigation, developer productivity assistance, translation of technical materials, search and retrieval functions and similar development support activities.
Such activities may be performed using Approved AI Services.
4. Retrieval-Augmented Generation
Licensee may use Covered Materials in RAG systems.
Licensee may create search indexes, vector databases, embedding stores, retrieval databases and similar retrieval mechanisms for the sole purpose of retrieving Covered Materials and supplying them as temporary context to an AI system.
Such retrieval mechanisms shall not be used for Model Training.
The existence of retrieval indexes or embeddings created solely for retrieval shall not, by itself, constitute Model Training.
5. AI Outputs
Any AI-generated output that contains, reproduces, reveals, describes, reflects or derives from Covered Materials shall be treated as Covered Materials to the extent necessary to preserve CSAB's confidentiality, intellectual property and trade secret rights.
Licensee shall not disclose such outputs except as otherwise permitted by the CETDL.
6. Private AI Model Training
Licensee shall not perform Model Training using Covered Materials without:
Such agreement may address permitted scope, ownership, confidentiality, audit rights, deployment restrictions, retraining restrictions and security requirements.
No approval shall be implied from silence, prior dealings, source code access or general permission to use AI Assistance.
7. Public AI Model Prohibition
Under no circumstances may Covered Materials be used to:
This prohibition applies whether the activity is performed directly or indirectly by Licensee or by a third party acting on Licensee's behalf.
8. Confidentiality and Security
Licensee shall use commercially reasonable efforts to utilize Private AI Environments whenever Covered Materials are used with artificial intelligence systems.
Licensee shall implement commercially reasonable technical and organizational safeguards designed to prevent unauthorized disclosure, unauthorized access, unauthorized retention, unauthorized publication, unauthorized model training and unauthorized use of Covered Materials.
Nothing in this Appendix reduces Licensee's obligations under Sections 2.7 or 6 of the CETDL.
9. Reservation of Rights
Except for the limited rights expressly granted in this Appendix, no right is granted to train AI models using Covered Materials, create derivative AI models, commercialize AI models derived from Covered Materials or use Covered Materials in any manner not specifically authorized by the CETDL or this Appendix.
All such rights remain expressly reserved by CSAB.
10. Survival
Sections 5, 6, 7, 8 and 9 of this Appendix shall survive termination of the CETDL.
Approved AI Services Schedule
1. General Rule
Only Approved AI Services may be used with Covered Materials.
Only enterprise, commercial or business-grade offerings may be used.
Consumer, personal, free-tier or publicly shared AI services are not Approved AI Services unless expressly identified below.
2. Approved AI Services
OpenAI
The following OpenAI offerings are approved:
- ChatGPT Enterprise
- ChatGPT Business (formerly Team)
- OpenAI API under a commercial agreement
provided that:
- CSAB source code is not used for Public AI Model training;
- business data protection controls remain enabled; and
- use remains consistent with this Appendix.
Microsoft
The following Microsoft offerings are approved:
- Microsoft 365 Copilot
- Microsoft Copilot for Microsoft 365
- GitHub Copilot Business
- GitHub Copilot Enterprise
- Azure OpenAI Service
provided such services remain subject to Microsoft's commercial data protection commitments.
Anthropic
The following Anthropic offerings are approved:
- Claude Team
- Claude Enterprise
- Claude API under a commercial agreement
The following Google offerings are approved:
- Gemini Enterprise
- Gemini Business
- Vertex AI













