Cross-corpus research synthesis
Open-Weight Frontier Release
Governance posture toward releasing frontier model weights publicly (Meta Llama, Mistral, DeepSeek vs. closed-weight Anthropic / OpenAI / DeepMind). EU AIA Recital 102 + Art. 53(2) carve-outs; CA SB-1047's failed framework; Meta Frontier AI Framework's explicit defence; emerging US export-control overlay.
Synthesised deterministically from 13 articles that engage this theme. Empirical consensus: contested · contested: Should frontier weight-release be governed by capability-tier (block above threshold) or by safety-evaluation-evidence (allow with pre-release red-team) or by recipient-restriction (export controls)? Three distinct frames currently in active conflict.. Full theme article: /wiki/open-weight-release. Machine-readable: /wiki/synthesis.json.
Cross-jurisdiction stances (4 govern, 12 engage)
| Instrument | Verdict | Provision excerpt / citation |
|---|---|---|
| EU AI Act | governs | The obligations in paragraph 1, points (a) and (b), shall not apply to providers of AI models released under a free and open-source licence … unless they are general-purpose AI models with systemic risks. (paraphrase) Art. 53(2) + Recital 102/104 — explicit open-source GPAI exemption (with caveats for systemic-risk models) |
| Executive Order 14110 on Safe, Secure, Trustworthy AI | implicit | §4.6 NTIA report on dual-use foundation models specifically addresses open-weight risk; not binding obligation |
| Interim Measures for Generative AI Service Management | implicit | Art. 8 — registration / safety assessment applies regardless of weight release modality |
| Seoul Declaration on Safe, Innovative and Inclusive AI | implicit | Frontier AI Safety Commitments apply to all 16 signatories regardless of open/closed weight stance (Meta is signatory) |
| California SB-1047: Safe and Secure Innovation for Frontier AI Models Act | governs | Vetoed bill — would have required covered models (incl. open-weight releases) to adopt a safety & security protocol + self-certified compliance, with independent third-party audits from 2026 (Anthropic + Meta objected on different grounds) |
| African Union Continental AI Strategy | implicit | Continental strategy frames AI capacity-building — open access to weights aligns with capacity goals |
| Anthropic Responsible Scaling Policy (RSP) v2 | implicit | RSP applies to Anthropic's models which are closed-weight; framework does not address third-party open release |
| OpenAI Preparedness Framework | implicit | Framework applies to OpenAI deployments (closed-weight); does not address third-party open release |
| Google DeepMind Frontier Safety Framework | implicit | Framework applies to Google DeepMind deployments (mostly closed); third-party open release not addressed |
| Meta Frontier AI Framework | governs | Framework's distinctive feature — explicit defence of open-weight release as governance posture; halt-training commitment if 'critical risk' threshold reached without mitigations |
| California SB 942: AI Transparency Act | governs | “If a covered provider licenses its GenAI system to a third party, the covered provider shall require by contract that the licensee maintain the system's capability to include a disclosure required by subdivision (b) in content the system creates or alters.” Cal. Bus. & Prof. Code § 22757.3(c) (added by SB 942, operative Aug. 2, 2026) — a covered provider that LICENSES its GenAI system to a third party must require by contract that the licensee preserve the § 22757.3(b) disclosure capability, and must revoke the license within 96 hours if the licensee disables it; reinforced by § 22757.3.2 (added by AB 853, operative Jan. 1, 2027), which bars a GenAI hosting platform distributing a system's source code or model weights from knowingly hosting a non-disclosing system |
| UN Global Digital Compact | implicit | “increase access to resources including open artificial intelligence models and systems, open training data and compute.” GDC Objective 5 capacity-building partnerships (A/RES/79/1, Annex I) |
Evidence convergence
Sources the corpus cites for this theme across multiple articles — a scientometric consensus signal computed from inline prose citations (the more articles independently cite a source, the more load-bearing it is for this theme). 20 sources are cited by ≥2 articles.
- 9×An interdisciplinary account of the terminological choices by EU policymakers ahead of the final agreement on the AI Act: AI system, general purpose AI system, foundation model, and generative AI — cited by 9 articles
- 7×Artificial intelligence and synthetic biology: biosecurity risks, dual-use concerns, and governance pathways — cited by 7 articles
- 7×The EU model of AI governance: regulating artificial intelligence through law and policy — cited by 7 articles
- 6×Two types of AI existential risk: decisive and accumulative — cited by 6 articles
- 6×Defending Compute Thresholds Against Legal Loopholes — cited by 6 articles
- 5×Infrastructure for AI Agents — cited by 5 articles
- 5×Missing the Mark: Adoption of Watermarking for Generative AI Systems in Practice and Implications Under the New EU AI Act — cited by 5 articles
- 5×Governing AI Agents — cited by 5 articles
- 5×International Agreements on AI Safety: Review and Recommendations for a Conditional AI Safety Treaty — cited by 5 articles
- 4×Multi-Agent Risks from Advanced AI — cited by 4 articles
- 3×Generative AI and data protection — cited by 3 articles
- 3×AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents — cited by 3 articles
- 3×China's semiconductor conundrum: understanding US export controls and their efficacy — cited by 3 articles
- 2×Audio deepfakes and the regulation of the landlords of creativity — cited by 2 articles
- 2×Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law (Council Eur.) — with Introductory Note — cited by 2 articles
- 2×Artificial Intelligence and Nuclear Weapons Proliferation: The Technological Arms Race for (In)visibility — cited by 2 articles
- 2×Large language models reflect the ideology of their creators — cited by 2 articles
- 2×The establishment of an international AI agency: an applied solution to global AI governance — cited by 2 articles
- 2×Navigating China's regulatory approach to generative artificial intelligence and large language models — cited by 2 articles
- 2×On the Societal Impact of Open Foundation Models — cited by 2 articles