# Frontier Model Safety Frameworks Double: More Companies Publishing AI Safety Standards

> There is a genuinely encouraging trend buried in the International AI Safety Report 2026: the number of companies publishing formal Frontier AI Safety Frameworks has more than doubled since 2025. On paper, that means more labs are documenting how they train models to refuse harmful requests, watermark AI-generated content, and define the risk thresholds at which they would pause a deployment. The report is careful not to oversell it, and that honesty is the useful part. Publishing a framework is not the same as being safe, and the authors flag that sophisticated attackers can still bypass current defenses, while the real-world effectiveness of these measures remains largely untested. The deeper worry is a pacing problem: capability gains keep widening the pathways to harm faster than anyone can measure actual misuse in the wild. So the doubling of frameworks is best read as a maturing of norms, an industry agreeing on what it should be disclosing, rather than proof that the systems are actually harder to misuse. It is progress on transparency, which is not nothing, and not yet progress on protection.

_Section: [Daily AI Updates](https://www.wortins.com/daily-ai) · Source: Inside Privacy · Published Tuesday, July 14, 2026_

## Wortins' read

There is a genuinely encouraging trend buried in the International AI Safety Report 2026: the number of companies publishing formal Frontier AI Safety Frameworks has more than doubled since 2025. On paper, that means more labs are documenting how they train models to refuse harmful requests, watermark AI-generated content, and define the risk thresholds at which they would pause a deployment. The report is careful not to oversell it, and that honesty is the useful part. Publishing a framework is not the same as being safe, and the authors flag that sophisticated attackers can still bypass current defenses, while the real-world effectiveness of these measures remains largely untested. The deeper worry is a pacing problem: capability gains keep widening the pathways to harm faster than anyone can measure actual misuse in the wild. So the doubling of frameworks is best read as a maturing of norms, an industry agreeing on what it should be disclosing, rather than proof that the systems are actually harder to misuse. It is progress on transparency, which is not nothing, and not yet progress on protection.

## Source

[Read the full story at Inside Privacy](https://www.insideprivacy.com/artificial-intelligence/international-ai-safety-report-2026-examines-ai-capabilities-risks-and-safeguards/)

## Related coverage

- [Europe faces AI talent exodus to US but emerging 'brain regain' momentum counters trend](https://www.wortins.com/story/europe-faces-ai-talent-exodus-to-us-but-emerging-brain-regai-7bddcc1e) — [Euronews](https://www.euronews.com/my-europe/2026/01/29/the-brain-drain-why-europe-cant-keep-the-talent-it-trains)
- [xAI launches Grok 4.5 specializing in coding, agents, and knowledge work](https://www.wortins.com/story/xai-launches-grok-4-5-specializing-in-coding-agents-and-know-8246856d) — [xAI](https://x.ai/news/series-e)
- [South Korea Commits $880 Billion to AI and Semiconductor Infrastructure Over 10 Years](https://www.wortins.com/story/south-korea-commits-880-billion-to-ai-and-semiconductor-infr-ec7afe2e) — [techspot.com](https://www.techspot.com/news/112926-south-korea-betting-880-billion-next-ai-race.html)
- [NVIDIA and ServiceNow Launch Project Arc: Persistent Autonomous Desktop Agent](https://www.wortins.com/story/nvidia-and-servicenow-launch-project-arc-persistent-autonomo-8e9f78ab) — [servicenow.com](https://newsroom.servicenow.com/press-releases/details/2026/ServiceNow-extends-agentic-AI-governance-from-desktops-to-data-centers-with-NVIDIA/default.aspx)
- [Scaled Cognition lands $100 million Series A for enterprise AI reliability](https://www.wortins.com/story/scaled-cognition-lands-100-million-series-a-for-enterprise-a-c0558e53) — [HPCwire](https://www.hpcwire.com/aiwire/2026/07/01/scaled-cognition-lands-100m-series-a-to-scale-reliable-enterprise-ai/)
- [Perplexity launches SPACE secure sandbox to let AI agents run safely at scale](https://www.wortins.com/story/perplexity-launches-space-secure-sandbox-to-let-ai-agents-ru-b2bfbad8) — [SiliconANGLE](https://siliconangle.com/2026/07/15/perplexity-launches-secure-sandbox-make-ai-agents-secure-powerful/)

---

_Curated and written by [Wortins](https://www.wortins.com) — The daily AI briefing. Every story links to its original source; the "Wortins read" on each is our own original analysis. [About Wortins & our editorial approach](https://www.wortins.com/about)._