# TechCrunch — OpenAI says it slowed Astra model development over security concerns

- Company: TechCrunch (techcrunch.com)
- Announced: 2026-08-07T22:48:24+00:00
- Category: safety-policy-update
- Subject: Reporting
- Models affected: Astra
- Coverage: Covered by 25 outlets
- Source: https://techcrunch.com/2026/08/07/openai-says-it-slowed-astra-model-development-over-security-concerns/
- Record: https://forck.live/items/3686-openai-says-it-slowed-astra-model-development-over-security-concerns

OpenAI announced it has suspended work on some aspects of its upcoming model Astra after an internal review found it reached a 'critical cybersecurity threshold,' meaning it could independently identify and carry out cyberattacks against well-protected real-world systems. The company is enacting stricter security controls and pausing internal activities involving Astra that don't meet these guardrails, while working with government agencies and AI safety organizations to test the model's capabilities.

## Evidence

Verbatim from https://techcrunch.com/2026/08/07/openai-says-it-slowed-astra-model-development-over-security-concerns/:

> OpenAI said Friday it has suspended work on some aspects of its upcoming model Astra after an internal review found it had made significant advancements in agentic coding and cybersecurity — enough to warrant concern over its capabilities.

## Around this story

Outlets this record can name and link:

- Bloomberg: https://bloomberg.com/news/articles/2026-08-07/openai-pauses-some-work-on-new-astra-model-over-cyber-concerns
- Reuters: https://reuters.com/legal/litigation/openai-flags-possible-critical-cybersecurity-risk-upcoming-model-tightens-2026-08-07
- Wall Street Journal: https://wsj.com/tech/ai/openai-pauses-some-work-on-new-ai-model-over-cybersecurity-concerns-8473a86f?st=V3qne7&reflink=desktopwebshare_permalink
- axios.com: https://axios.com/2026/08/07/openai-astra-model-delay-cybersecurity-risks
- The Verge — OpenAI puts the brakes on a new model because it’s supposedly too powerful: https://theverge.com/ai-artificial-intelligence/976948/openai-astra-model-pause-critical-cyber-capabilities
- The Register — OpenAI pledges to add Astra security as Anthropic loosens Fable's leash: https://theregister.com/ai-and-ml/2026/08/08/openai-pledges-to-add-astra-security-as-anthropic-loosens-fables-leash/5285161
- The Decoder — OpenAI flags its new Astra model as potentially reaching the highest cybersecurity risk level for the first time: https://the-decoder.com/openai-flags-its-new-astra-model-as-potentially-reaching-the-highest-cybersecurity-risk-level-for-the-first-time
- SiliconANGLE — OpenAI reveals upcoming Astra model may possess 'critical’ hacking capabilities - SiliconANGLE: https://siliconangle.com/2026/08/07/openai-reveals-upcoming-astra-model-may-possess-critical-hacking-capabilities
- Reuters — OpenAI flags possible critical cybersecurity risk in upcoming model, tightens controls - The Economic Times: https://economictimes.indiatimes.com/tech/artificial-intelligence/openai-flags-possible-critical-cybersecurity-risk-in-upcoming-model-tightens-controls/articleshow/133040002.cms
- The New Stack — The AI model OpenAI won't release yet — and what it found in testing: https://thenewstack.io/openai-astra-cybersecurity-delay
- Moneycontrol — OpenAI strengthens security controls as upcoming Astra AI model nears critical cyber threshold- Moneycontrol.com: https://moneycontrol.com/technology/openai-strengthens-security-controls-as-upcoming-astra-ai-model-nears-critical-cyber-threshold-article-13998497.html
- MacRumors — OpenAI Delays Next Major AI Model 'Astra' Over Critical Hacking Concerns: https://macrumors.com/2026/08/07/openai-astra-model-hacking-concerns
- PCWorld — OpenAI pumps the brakes on new Astra model over cybersecurity concerns: https://pcworld.com/article/3208734
- TestingCatalog AI News — OpenAI says Astra may have reached Critical cyber threshold: https://testingcatalog.com/openai-says-astra-may-have-reached-critical-cyber-threshold
- Bloomberg Law — OpenAI Pauses Some Work on New Astra Model on Cyber Concerns (1): https://news.bloomberglaw.com/tech-and-telecom-law/openai-pauses-some-work-on-new-astra-model-on-cyber-concerns-1
- RuntimeWire — OpenAI slows Astra development after cyber tests trigger its highest risk threshold: https://runtimewire.com/article/openai-slows-astra-development-critical-cyber-risk
- RuntimeWire — OpenAI slows Astra release after cyber tests raise critical-risk concerns: https://runtimewire.com/article/openai-slows-astra-release-cyber-risk
- Blockonomi — OpenAI Halts Astra AI Development Over Autonomous Hacking Concerns - Blockonomi: https://blockonomi.com/openai-halts-astra-ai-development-over-autonomous-hacking-concerns
- Yahoo Finance — OpenAI says its upcoming Astra model may have 'critical' cybersecurity capabilities amid rash of AI model hacks: https://finance.yahoo.com/technology/article/openai-says-its-upcoming-astra-model-may-have-critical-cybersecurity-capabilities-amid-rash-of-ai-model-hacks-194909085.html
- The Mac Observer: https://macobserver.com/news/openai-pauses-astra-software-release-over-severe-hacking-risks
- Business Standard: https://business-standard.com/technology/tech-news/openai-pauses-work-on-new-astra-model-to-boost-safeguards-over-cyber-risks-126080800089_1.html
- Tech Times — OpenAI Pauses Astra After Tests Reveal Autonomous Zero-Day Exploit of Hardened Systems: https://techtimes.com/articles/323628/20260808/openai-pauses-astra-after-tests-reveal-autonomous-zero-day-exploit-hardened-systems.htm
- Cyber Security News — OpenAI Slows Down New Astra Model Development to Measure Cybersecurity Capabilities: https://cybersecuritynews.com/openai-slows-down-new-astra-model
- Seoul Economic Daily — OpenAI Delays 'Astra' Launch Over Hacking Concerns: https://en.sedaily.com/international/2026/08/08/openai-delays-astra-launch-over-hacking-concerns
- Forkast — OpenAI Pauses Astra at ‘Critical’ Cyber Threshold — First Frontier Model to Trigger Highest Preparedness Framework Level: https://forkast.news/openai-pauses-astra-at-critical-cyber-threshold-first-frontier-model-to-trigger-highest-preparedness-framework-level
- OpenAI: https://openai.com/index/responding-next-frontier-critical-cyber-capabilities
- Finimize: https://finimize.com/content/openai-put-its-astra-model-through-extra-cyber-safeguards
- Nairametrics — OpenAI flags critical cybersecurity risk in AI model weeks after ‘Hugging Face incident’: https://nairametrics.com/2026/08/07/openai-flags-critical-cybersecurity-risk-in-ai-model-weeks-after-hugging-face-incident
- Sources — OpenAI's Astra pause: https://sources.news/p/openai-astra-pause

---

Record: https://forck.live/items/3686-openai-says-it-slowed-astra-model-development-over-security-concerns
Catalogue: https://forck.live/llms.txt
Feed: https://forck.live/feed.md