The Jerusalem PostTV TIME: Best intense thrillers while awaiting ‘Fauda’וואלהרוסיה דורשת הבהרות מארה"ב וטורקיה על העברת נשק אפשרית לאוקראינהRTP DesportoEuropeus de atletismo. Vitória Oliveira bate recorde na meia maratona em marchaESPN DeportesDiana Flores dedica el Mundial a Julio Ocaña, el coach que considera su mentorInquirer‘Parents responsible for protecting kids online, not just government’RTL BoulevardTurnster Veerman mist nipt brugmedaille bij EK, Visser laatsteMint‘Biggest scam ever in the world’: Nithin Kamath asks startup founders not to brag about AI; social media reactsMundo DeportivoCity 0-2 Barça: Vicky marca el segundoSportstarHockey World Cup: Skipper Harmanpreet hails India’s 3-1 win over Wales in openerRai NewsI servizi russi arrestano una donna: "Ha messo la bomba che ha ucciso un ufficiale a Sebastopoli"MyJoyOnlinePhotos: Police engage religious leaders at National Police HeadquartersBBC News BrasilTerremoto de magnitude 7,7 mata ao menos 47 pessoas na Indonésia
The Daily Newsstand · Free, Always
Saturday, August 15, 2026
Running 2 days

OpenAI Slows Development of Astra Model

OpenAI has paused the development of its new AI model, Astra, due to concerns that it could autonomously perform cyberattacks. The company stated that preliminary evaluations indicated the model might possess critical cybersecurity capabilities.

Where they agreeOpenAI has slowed the development of its Astra model following internal assessments regarding its potential cybersecurity risks.

Where they differThe French report links the decision to broader industry concerns about autonomous agents, while the Spanish and Arabic reports focus specifically on the internal safety framework.

Key playersSam Altman (CEO)

Ask about this story

AI — answers come only from the articles on this page, and can be wrong. When they don’t contain the answer, it says so instead of guessing.

SpainFranceInternational

Monday, August 10Latest

Earlier

Saturday, August 8

OpenAI has paused work on its AI model Astra due to security concerns regarding its advanced cyber capabilities. The company stated it will implement stricter testing and safety protocols before proceeding.

Stories are grouped automatically, headlines are machine-translated, and the summary, the agree/differ lines and the assistant’s answers are all written by AI from the articles themselves. Any of it can be wrong. The headline as printed and a link to the publisher are shown throughout so you can check everything against the source.