Skip to content
Säkerhet· NewsNot in EU

OpenAI pauses work on AI model Astra due to safety concerns

OpenAI is pausing parts of the development of its upcoming AI model, Astra. Internal testing indicates the model may have reached the highest risk level for cyber capabilities.

By the Aheadline editorial team·12 aug. 2026·2 min read·Source: Entity-watch: AI SverigeVerifierad signalAI-generated
OpenAI pauses work on AI model Astra due to safety concerns
OpenAI pauses work on AI model Astra due to safety concerns
OpenAI pauses work on AI model Astra due to safety concerns
By · Policy- & EU-reporter
Last updated

What happened?

OpenAI has decided to pause parts of the development and restrict internal access to its upcoming AI model, Astra. The decision follows internal evaluations showing that the model exhibits advanced cyber capabilities. The company states it cannot rule out that Astra has reached the 'Critical' level according to its internal Preparedness Framework.

Key facts

ModellnamnAstra
Identifierad risknivåCritical (Preparedness Framework)
Rapporterat datum10 augusti 2026

Why it matters

For a model to be classified at the 'Critical' level, it must demonstrate the ability to autonomously identify and exploit zero-day vulnerabilities, as well as plan and execute complex cyberattacks without human intervention. The identification of these capabilities has prompted enhanced monitoring and stronger protections for the model weights to prevent misuse.

Who is affected?

The decision primarily affects OpenAI's own developers and researchers, whose internal tool and network access is being restricted. In the longer term, external systems integrators and security researchers awaiting the rollout of next-generation AI models will also be affected.

Impact on the EU

The decision reflects global security requirements and existing frameworks regarding advanced AI. The measures apply generally to the company's global infrastructure, thereby impacting the handling of the model within the EU and Sweden as well.

What else you should know

The information is based on OpenAI's own Preparedness Framework, which categorizes risks into four levels ranging from 'Low' to 'Critical'. Reaching the highest risk level automatically triggers internal protocols to prevent the dissemination of potentially harmful model weights.

Frequently asked questions

Quick answers about this story

Vad har hänt?
OpenAI har pausat delar av utvecklingen samt den interna användningen av AI-modellen Astra på grund av säkerhetsrisker kopplade till avancerade cyberförmågor.
När hände det?
Beskedet rapporterades den 10 augusti 2026 efter interna säkerhetstester hos OpenAI.
Varför spelar det roll?
Händelsen markerar att AI-modeller kan ha nått nivån Critical i OpenAI:s ramverk, vilket innebär teoretisk förmåga att autonomt upptäcka zero day-sårbarheter och utföra cyberattacker.
Hur påverkar det framtida lanseringar?
Säkerhetsstoppet innebär att framtida lanseringar fördröjs tills riskerna har utvärderats och hanterats enligt gällande säkerhetsprotokoll.
Original source
Entity-watch: AI Sverige·computersweden.se

The link opens in a new window and leads to the publisher's own site.

Verifierad signal

Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.

AI-verktyg i artikeln

Topics

#OpenAI
[ STAY UP TO DATE ]

Get similar news straight to your inbox

No affiliate linksCancel anytimeGDPR-friendly
[ Frequency ]
[ What do you want to read about? ]

You'll receive updates on 2 topics.

The reader's room

Send in a question or an addition. The newsroom reads everything before it's published and replies when relevant. No AI-generated text – just people.

Sign in to submit a comment or question.

Loading comments…
How this affects you

Read the article through your role

  • Decide whether this affects strategy over 6–12 months or is just noise.
  • Discuss with leadership: do we own the right question or does ownership need to move?
  • Ask: what risk are we taking by NOT acting on this this quarter?

Generated angle — not editorial analysis of "OpenAI pauses work on AI model Astra due to safety concerns"