GPAI Model Provider

stable gpai-model-provider · v1.1

A provider that places a general-purpose AI model on the market and carries the Chapter V duties that come with it: model documentation and downstream information, a copyright policy and rights reservation compliance, the public summary of training content and, once a model is designated as carrying systemic risk, a systemic risk framework and acceptance determination, evaluation under standardised protocols, protection of the model weights and infrastructure and serious incident reporting to the AI Office, with the Safety and Security Model Report recommended. The statement is seeded from saas-ai-provider because an organisation that trains and offers a model runs the same security, privacy and governance programme, and re-judged where the seat differs: the high-risk conformity rows are out of scope for a model as such and the Art.54 authorised representative arm is conditional on establishment outside the Union. A provider that also serves its model through a product holds the provider role and saas-ai-provider alongside this profile (ADR-046).

Roles: GPAI Model Provider
Deployment models: Cloud, multi-tenant SaaSCloud, single tenantOn premisesHybridEdgeOn deviceEmbeddedAir gapped
Risk classes: gpaigpai-systemic
Jurisdictions: EU
Frameworks: SOC2ISO-27001NIST-800-53CSA-CCMCSA-AICMISO-42001NIST-AI-RMFNIST-AI-600-1OWASP-LLMOWASP-AGENTICEU-AI-ActGPAI-COPGDPR

Applicability statement

201 controls Download CSV

required 178 controls In scope for this profile. The control has to be in place and evidenced.

GOV · Governance & Risk

GOV-001 Information Security Policy core

GOV-002 Information Security Roles and Responsibilities core

GOV-003 Management Commitment and Accountability core

GOV-004 Information Security Programme core

GOV-005 Risk Assessment core

GOV-006 Risk Management Programme core

GOV-007 Risk Treatment and Remediation Tracking core

GOV-008 Fraud Risk Assessment core

GOV-009 Segregation of Duties core

GOV-010 Legal, Regulatory and Contractual Compliance Inventory core

GOV-011 Compliance Monitoring and Internal Audit core

GOV-012 Continuous Monitoring Strategy core

GOV-013 Exception and Requirement Determination Register core

GOV-014 Asset Inventory core

GOV-015 Intellectual Property Rights Management core

GOV-016 Records and Information Governance core

GOV-018 Threat Intelligence Programme core

GOV-019 Information Security in Project Management core

GOV-020 Independent Security Review core

GOV-021 Audit and Assurance Policy core

GOV-022 Privacy Programme and Data Protection Policy core

GOV-023 Security Measures Performance Measurement core

GOV-024 Documented Operating Procedures core

GOV-028 Regulatory Cooperation and Supervisory Access role duty

The provider seat is what brings a customer's supervisor to the organisation: the duty arrives through a customer contract rather than through a law binding the organisation directly. A provider whose customers are unregulated still holds the row, because what is tested is that the route exists, that the access is committed and that no contract term obstructs it. The register may legitimately be empty for a period, in which case the route is tested instead.

IAM · Identity & Access Management

IAM-001 Access Control Policy core

IAM-002 Identity Inventory and Unique Identifiers core

IAM-003 User Account Lifecycle Management core

IAM-004 Access Review and Recertification core

IAM-005 Least Privilege and Need-to-Know Enforcement core

IAM-007 Privileged Access Management core

IAM-008 Multi-Factor Authentication core

IAM-009 Authentication Information Management core

IAM-010 Service Account and Non-Human Identity Management core

IAM-011 Remote Access Controls core

IAM-012 Session Management core

IAM-013 Logon Failure and Account Lockout core

IAM-014 Access to Source Code and Development Assets core

IAM-015 Role-Based Access Control core

IAM-016 Dedicated Administration Systems core

DAT · Data Protection

DAT-001 Data Classification Scheme core

DAT-002 Information Labelling core

DAT-003 Encryption at Rest core

DAT-004 Encryption in Transit core

DAT-005 Cryptographic Key Management core

DAT-006 Data Inventory and Records of Processing core

DAT-007 Data Minimisation and Purpose Limitation core

DAT-008 Data Retention and Deletion core

DAT-009 Privacy Notice and Transparency core

DAT-010 Consent Management core

DAT-011 Data Subject Rights Fulfilment core

DAT-012 Data Protection by Design and Default core

DAT-013 Data Protection Impact Assessment core

DAT-015 Data Transfer Controls core

DAT-016 Data Masking and Pseudonymisation core

DAT-017 Data Leakage Prevention core

DAT-018 Data Protection Officer core

DAT-019 Lawful Basis for Processing core

DAT-020 Accuracy of Personal Data core

DAT-024 Special Category and Criminal Conviction Data core

DAT-025 Automated Decision-Making and Profiling Rights core

APP · Application Security

INF · Infrastructure & Cloud Security

MON · Monitoring & Logging

MON-001 Audit Log Scope and Generation core

MON-002 Log Integrity and Protection core

MON-003 Log Retention core

MON-004 Centralised Log Management core

MON-005 Security Monitoring and Alerting core

MON-008 Detection Content Lifecycle core

MON-009 Log Access Control and Sensitive Data in Logs core

VND · Vendor & Third-Party Risk

VND-001 Vendor Risk Assessment and Due Diligence core

VND-002 Security Requirements in Vendor Contracts core

VND-003 Sub-Processor Management core

VND-004 Cloud Service Provider Security Management core

VND-006 Vendor Monitoring and Performance Review core

VND-007 Vendor Access Controls core

VND-008 Vendor Offboarding core

VND-010 Third-Party Data Disclosure Controls core

BCM · Business Continuity

AIG · AI Governance

AIG-001 AI Policy core

AIG-002 AI Roles and Responsibilities core

AIG-003 AI System Inventory core

The inventory holds each model with its designation status. The Art.52(1) notification to the Commission within two weeks of meeting the systemic-risk threshold is the threshold notification clause of this control.

AIG-005 AI Risk Management Process core

AIG-006 AI Impact Assessment core

AIG-007 AI System Requirements and Design Documentation core

AIG-008 AI System Verification, Validation and Testing core

AIG-009 AI System Deployment and Change Management core

AIG-010 AI Model Registry and Versioning core

AIG-011 AI System Decommissioning core

AIG-012 Training Data Management and Quality core

AIG-013 Training Data Provenance core

AIG-014 Special Category Data in Training and Evaluation Datasets core

AIG-015 AI System Technical Documentation core

AIG-016 AI Interaction and Output Disclosure core

Art.50 disclosure binds every provider of an interacting or generating system, whatever its risk class.

AIG-017 AI Model Explainability core

AIG-018 AI System Operational Monitoring core

AIG-019 AI Model Performance and Drift Detection core

AIG-020 AI System Event Logging core

AIG-021 AI Incident Response and Error Communication core

AIG-022 Human Oversight of AI Outputs core

AIG-023 AI System Override and Safe-State Mechanisms core

AIG-024 Prohibited AI Practices core

Art.5 binds every provider, whatever the risk class of its other systems.

AIG-025 AI Fairness and Bias Controls core

AIG-026 AI Security and Adversarial Robustness core

AIG-027 AI Output Validation and Confidence Controls core

AIG-028 Hallucination and Factual Accuracy Controls core

AIG-029 Prompt Injection Protection core

AIG-031 AI Misuse, Jailbreak and Abuse Detection core

AIG-032 Third-Party AI Risk Management core

AIG-034 Customer and Deployer Obligations Communication role duty

Instructions for use, the provider identification and the customer agreement are provider duties (Art.13, Art.16.2, ISO 42001 A.8).

AIG-038 AI Feedback and Adverse Impact Reporting Channel core

AIG-039 Responsible and Intended Use of AI Systems core

The provider is also a user of AI systems, including the models beneath its product. The use register covers both.

AIG-040 Customer Data Use in Model Training role duty

The statement on customer data in training and the opt-out are the provider's to give.

AIG-042 Agentic Tool Permissions and Action Authorisation core

AIG-043 AI Non-Conformity Corrective Action and Authority Cooperation role duty

Art.53(3) puts a cooperation duty on a general-purpose model provider: responding to information requests and taking part in an authority-led investigation. The request register here serves it; the mapping note records that the investigation half is not written into the control.

AIG-047 General-Purpose Model Documentation and Downstream Information role duty

Art.53(1)(a) and (b): the model documentation and the downstream information, for every general-purpose model placed on the market. Art.53(2) relieves a free and open-source model without systemic risk of both; the relief is recorded against the inventory entry.

AIG-048 Public Training Content Summary role duty

Art.53(1)(d): the public summary of training content, for every general-purpose model, with no open-source relief.

AIG-049 Copyright Policy and Rights Reservation Compliance role duty

Art.53(1)(c): the copyright policy and rights reservation compliance, for every general-purpose model, with no open-source relief.

AIG-055 Training Pipeline Security core

The training pipeline is the model providers own system.

AIG-056 Agent Memory and Context Integrity core

AIG-057 Inbound Synthetic Media Detection core

Required, as the library states every control whose trigger is a feature set rather than a facet (DAT-018, AIG-014, AIG-042, AIG-056): the documented assessment the control opens with is the first artefact. It bites where a system relies on user-supplied or external audio, image or video as evidence of a person's identity, of the provenance of an artefact or of a real-world event, which is a live case for identity verification, onboarding, claims and content-moderation features. An organisation whose assessment finds no such intake point records that finding and the detection limb has nothing to attach to. Restated from recommended on the S13 benchmark (ADR-049 amendment, 2026-09-16).

HRS · Human Resources Security

conditional 17 controls In scope when the stated condition holds. Read the condition before deciding.