GAISSF / D1
Model Integrity & Adversarial Robustness
Domain purpose
Controls and requirements
D1-CTL-01Dataset Provenance & Poisoning Prevention
Protect training investment ($50k-$500k per model) from backdoored data.
D1-CTL-02Model Extraction Resistance
Protect $5M+ model IP from theft via API.
D1-CTL-03Behavioral Drift Detection
Prevent undetected model degradation causing business loss.
D1-CTL-04Federated Learning Poisoning Prevention
Protect multi-party models from malicious clients.
D1-CTL-05Embedding Space Robustness
Ensure semantic filters work under adversarial conditions.
D1-CTL-06Post-Quantum Model Signing & Crypto Hardening
Future-proof model supply chain against quantum attack.
D1-CTL-07Lora/Adapter Integrity Verification
Protect fine-tuning pipeline ($50k-$500k per model) from backdoored adapters.
D1-CTL-08Model Merge Attack Detection
Prevent safety-evasive merged models from entering production.
D1-CTL-09Quantization Backdoor Screening
Ensure quantization doesn't activate hidden backdoors.
Implementation use
Determine applicability using the framework scope and system context. Implementation should be proportionate to risk and supported by evidence sufficient to validate the intended outcome.