Google DeepMind Frontier Safety Framework is a Voluntary code from US, adopted on 2024-05-17 and effective 2024-05-17. Current status: In force. Critical Capability Levels (CCL) regime spanning autonomy, biosecurity, cybersecurity, and persuasion domains. Distinct vocabulary from Anthropic ASL + OpenAI Preparedness — designed for cross-domain elicitation; each CCL triggers domain-specific mitigations including model-weight access controls + enhanced red-teaming. Seoul Frontier AI Safety Commitments signatory. Alphabet-published; effective across Google DeepMind frontier-model releases.
Scope and obligations
Critical Capability Levels (CCL) regime spanning autonomy, biosecurity, cybersecurity, and persuasion domains. Distinct vocabulary from Anthropic ASL + OpenAI Preparedness — designed for cross-domain elicitation; each CCL triggers domain-specific mitigations including model-weight access controls + enhanced red-teaming. Seoul Frontier AI Safety Commitments signatory. Alphabet-published; effective across Google DeepMind frontier-model releases.
Google DeepMind Frontier Safety Framework addresses 2 contested AI-governance topics explicitly, 2 via general principles,.
Topics governed
- governsFoundation Models / GPAI— FSF applies to Google DeepMind frontier-model releases
- implicitTransparency Obligations— FSF publication discloses framework + thresholds; per-evaluation outputs not consistently public
- governsCatastrophic & Existential Risk— FSF Critical Capability Levels (CCL) — explicit thresholds for autonomy, biosecurity, cyber, persuasion
- implicitInternational Coordination— Seoul Frontier AI Safety Commitments signatory; UK AISI pre-deployment evaluation cooperation
Cross-jurisdiction comparison
How peer instruments treat the topics Google DeepMind Frontier Safety Framework governs.
| Topic | EU-AIA-2024 | US-EO-14110 | US-EO-14179 | UK-WHITEPAPER-2023 | CN-GENAI-2023 | G7-HIROSHIMA | OECD-AI-PRIN | COE-AI-CONV | UN-RES-2024 | NIST-AI-RMF | BLETCHLEY-2023 | SEOUL-2024 | NIST-AI-RMF-GENAI | CA-SB-1047 | IN-DPDP-2023 | BR-AIBILL-2024 | ASEAN-AI-GUIDE-2024 | AU-AI-STRATEGY-2024 | ANTHROPIC-RSP-2024 | OPENAI-PREPAREDNESS-2023 | META-FRONTIER-2024 | UK-US-AISI-MOU-2024 | WH-VOLUNTARY-2023 | SG-MODEL-AI-2024 | JP-METI-AI-2024 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Foundation Models / GPAI | governs | governs | silent | implicit | governs | governs | implicit | implicit | silent | governs | governs | governs | governs | governs | implicit | governs | implicit | silent | governs | governs | governs | governs | governs | governs | governs |
| Catastrophic & Existential Risk | implicit | governs | silent | implicit | silent | governs | silent | silent | implicit | implicit | governs | governs | governs | governs | silent | governs | silent | silent | governs | governs | governs | implicit | implicit | silent | silent |
How to cite this article
APA 7
Policy Window. (2024). Google DeepMind Frontier Safety Framework [Wiki article]. https://policywindow.org/wiki/deepmind-fsf
Chicago 17
Policy Window. 2024. "Google DeepMind Frontier Safety Framework." Wiki article. https://policywindow.org/wiki/deepmind-fsf.
BibTeX
@misc{policywindow-deepmind-fsf,
title = {Google DeepMind Frontier Safety Framework},
author = {Policy Window},
year = {2024},
howpublished = {Google DeepMind Frontier Safety Framework (May 2024)},
url = {https://policywindow.org/wiki/deepmind-fsf},
note = {Primary source: https://deepmind.google/discover/blog/introducing-the-frontier-safety-framework/}
}Related instruments
- Executive Order 14110 on Safe, Secure, Trustworthy AI · US
- G7 Hiroshima AI Process Code of Conduct · G7
- Bletchley Declaration on AI Safety · global
- Seoul Declaration on Safe, Innovative and Inclusive AI · global
- NIST AI RMF Generative AI Profile · US
- California SB-1047: Safe and Secure Innovation for Frontier AI Models Act · US
- Brazil AI Bill (PL 2338/2023) · BR
- Anthropic Responsible Scaling Policy (RSP) v2 · US
- OpenAI Preparedness Framework · US
- Meta Frontier AI Framework · US
References
- Google DeepMind Frontier Safety Framework (May 2024)
- FSF applies to Google DeepMind frontier-model releases
- FSF publication discloses framework + thresholds; per-evaluation outputs not consistently public
- FSF Critical Capability Levels (CCL) — explicit thresholds for autonomy, biosecurity, cyber, persuasion
- Seoul Frontier AI Safety Commitments signatory; UK AISI pre-deployment evaluation cooperation
Take this further — sign up free
Save, compare, or get alerts when Google DeepMind Frontier Safety Framework changes. Policy Window is the analyst workbench layered on top of this wiki — free for researchers, civil society, and verified policymakers.