# AI companies retreat from safety pledges as models grow more powerful

> A new Future of Life Institute report lands an uncomfortable verdict on the AI industry: even as models grow more capable, the companies building them are quietly walking back their safety commitments. In the group's scorecard, no major lab does well. Anthropic ranks first and still earns only a C+, while OpenAI and Google DeepMind each land at C. The specific finding that stands out is the retreat on danger thresholds. Several companies have weakened or removed earlier promises to pause development if a model crossed certain capability lines, exactly the guardrails that were meant to matter most as systems become more powerful. In other words, the commitments are loosening at the moment they would count. It is a reminder that voluntary self-governance tends to bend under competitive pressure. With frontier models advancing faster than any external oversight, reports like this are increasingly the only public accounting of whether the labs are keeping the promises they made when the stakes felt more hypothetical.

_Section: [Daily AI Updates](https://www.wortins.com/daily-ai) · Source: Axios · Published Saturday, July 18, 2026_

## Wortins' read

A new Future of Life Institute report lands an uncomfortable verdict on the AI industry: even as models grow more capable, the companies building them are quietly walking back their safety commitments. In the group's scorecard, no major lab does well. Anthropic ranks first and still earns only a C+, while OpenAI and Google DeepMind each land at C. The specific finding that stands out is the retreat on danger thresholds. Several companies have weakened or removed earlier promises to pause development if a model crossed certain capability lines, exactly the guardrails that were meant to matter most as systems become more powerful. In other words, the commitments are loosening at the moment they would count. It is a reminder that voluntary self-governance tends to bend under competitive pressure. With frontier models advancing faster than any external oversight, reports like this are increasingly the only public accounting of whether the labs are keeping the promises they made when the stakes felt more hypothetical.

## Source

[Read the full story at Axios](https://www.axios.com/2026/07/07/report-ai-safety-pledges)

## Related coverage

- [Anthropic Prepares for October IPO as Fable 5 Dominance Solidifies](https://www.wortins.com/story/anthropic-prepares-for-october-ipo-as-fable-5-dominance-soli-98c09b96) — [GuruFocus](https://www.gurufocus.com/news/8960658/anthropic-prepares-for-potential-ipo-outpacing-openai)
- [Suno AI Music Generator Faces Key Summary Judgment Hearing in Copyright Litigation](https://www.wortins.com/story/suno-ai-music-generator-faces-key-summary-judgment-hearing-i-6062f1bb) — [TLDL](https://www.tldl.io/resources/ai-music-generators-suno-udio-2026)
- [Apple Sues OpenAI, Apple's Real Problem](https://www.wortins.com/story/apple-sues-openai-apple-s-real-problem-e3136f7b) — [Stratechery](https://stratechery.com/2026/apple-sues-openai-apples-real-problem/)
- [Google's James Manyika: Betting Doomers Are Wrong About AI and Jobs](https://www.wortins.com/story/google-s-james-manyika-betting-doomers-are-wrong-about-ai-an-fb211e76) — [Platformer](https://www.platformer.news/james-manyika-google-ai-jobs-io-2026/)
- [Perplexity Launches SPACE Sandbox for Secure AI Agent Execution](https://www.wortins.com/story/perplexity-launches-space-sandbox-for-secure-ai-agent-execut-1fd03094) — [SiliconANGLE](https://siliconangle.com/2026/07/15/perplexity-launches-secure-sandbox-make-ai-agents-secure-powerful/)
- [Healthcare AI Diagnostics Adoption Accelerates Despite Safety Concerns](https://www.wortins.com/story/healthcare-ai-diagnostics-adoption-accelerates-despite-safet-5b905590) — [Radiology Business](https://radiologybusiness.com/topics/artificial-intelligence/navigating-ai-diagnostic-dilemma-healthcares-no-1-patient-safety-concern-2026)

---

_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)._