AI-as-a-Service security — model isolation + prompt-injection + AIBOM

Updated

AI-as-a-Service (per-API-call inference, fine-tuning offerings, RAG-as-a-service) has unique threats: tenant-prompt isolation, prompt-injection, training-data contamination, AIBOM transparency.

Top security risks

Prompt injection

Customer A's adversarial prompt modifies behavior for Customer B in shared-state systems.

Training-data contamination

Customer prompts/responses leak into your next fine-tune by mistake.

Cross-tenant prompt leakage

Shared model state lets one tenant's prompts or responses surface in another tenant's session.

Service-account key leakage

Customer's API key leaked → LLMjacking attack on your inference budget.

Regulatory context

Securie focuses on the security-engineering surface: tenant-isolated inference, prompt-injection defense, training-data opt-in, and per-tenant spend caps verified on every change.

Checklist

  • Tenant-isolated inference (no shared model state)
  • Prompt-injection defense (Llama Guard 4)
  • AIBOM published per /legal/ai-bill-of-materials
  • Training-data opt-in only
  • Per-tenant LLM spend caps
What your buyers look for

Buyers ask for AIBOM (CycloneDX 1.6), training-data practice, prompt-injection defense, and model-card documentation. Show a verifiable security posture for each.