Model distillation attack
Mass prompting extracts reasoning from a proprietary model
categoryAI/ML and emerging attacksseverityMediumlabnot yet measured
Classification
16 values: 5 imported, 1 derived, 10 machine-classified. What do these mean?
Exploitability
| Attack vector | AI model manipulation |
|---|---|
| Reach (CVSS) | Network |
| Attack complexity | Low |
| Privileges required | Low |
| User interaction | None |
| Attacker skill required | Medium |
| Resources required | High |
| Likelihood / prevalence | Medium |
| Maturity | Emerging |
Impact & scope
| Severity | Medium |
|---|---|
| Impact | ConfidentialityFinancial |
| Platform | AI System |
| Environment | AI system |
| Target asset class | ML model / agent |
| Kill-chain phase (tactic) | Exfiltration |
Detection & response
Mitigated by 4
- AML.M0002 atlas-mitigation
- AML.M0004 atlas-mitigation
- AML.M0024 atlas-mitigation
- AML.M0035 atlas-mitigation
Framework references
- Primary source: https://cloud.google.com/blog/topics/threat-intelligence/from-prompting-to-autonomy-the-evolution-of-adversarial-ai/
- atlas: AML.T0024.002 (primary)
Evidence & sources
- ev-0001 (Cyber Attack Taxonomy) — Name, definition and Example/ID taken verbatim from the 'AI/ML and emerging attacks' table, line 620 of Cyber-Attack-Taxonomy.md
- ev-0002 (Seeded-record enrichment run 2026-09-30 (crosswalk + facet classification)) — Crosswalk to atlas:AML.T0024.002 as the nearest framework entries; ids verified to exist in the cached catalogues.