Open-Source Model Risk

Risk Profile

Open-source models (Llama, Mistral, Mixtral, Falcon, etc.) offer control and cost advantages but introduce supply chain and operational risks.

Key Risks

Model Integrity

  • Pickle deserialization: Many model formats execute arbitrary code on load
  • Backdoored weights: Malicious models uploaded to public hubs pass benchmarks but contain hidden behaviors
  • Fine-tune poisoning: Community fine-tunes may include harmful training data

Operational Risk

  • No vendor support: You own the entire stack — inference, monitoring, patching
  • Security patches lag: Vulnerabilities in model serving software may not have rapid fixes
  • Talent dependency: Requires ML engineering expertise to operate

Compliance Risk

  • License confusion: Some "open" models have restrictive licenses (Llama's acceptable use policy)
  • Training data provenance: You may not know what data the model was trained on
  • Liability: No vendor to share liability if the model causes harm

Controls

ControlImplementation
Safetensors onlyOnly load models in safetensors format — no pickle execution risk
Hash verificationVerify model file hashes against published checksums
Model scanningScan model files for malicious payloads before loading
Sandboxed inferenceRun models in isolated containers with no network access to sensitive systems
License reviewLegal review of model license before deployment
Provenance documentationDocument model source, version, and modification history
Safety evaluationRun safety benchmarks before production deployment
Update processDefined process for updating model versions with testing gates