OpenAI says it slowed Astra model development over security concerns

Center-Left: <a href="https://www.axios.com/2026/08/06/ai-singularity-intelligence-explosion" rel="noopener nofollow">Welcome to the singularity: AI&#39;s architects say the next era of human history is here — www.axios.com
Center-Left: <a href="https://www.theguardian.com/technology/2026/aug/05/meta-ai-model-hack-training" rel="noopener nofollow">Meta says its AI model hacked into another company during testing — www.theguardian.com
Center-Left: <a href="https://www.theguardian.com/technology/2026/aug/05/openai-anthropic-models-went-rogue-cybersecurity-test-ai-security-institute" rel="noopener nofollow">AI models shock UK testers by using fake identities to try to trick developers — www.theguardian.com
Center-Right: <a href="https://www.cnbc.com/2026/08/06/ai-kill-switch-bill-openai-anthropic-meta.html" rel="noopener nofollow">&#39;AI Kill Switch&#39; bill needs to be passed this year amid ongoing rogue agent hacks, Rep. Lieu says — www.cnbc.com
Right: <a href="https://www.theblaze.com/tech/trump-writes-new-rules-to-stop-rogue-ai-but-they-re-classified" rel="noopener nofollow">Trump writes new rules to stop rogue AI — but they&#39;re classified — www.theblaze.com
Right: <a href="https://www.theblaze.com/tech/trump-bans-chinese-robots-but-your-robot-vacuum-may-be-spying-on-you-already" rel="noopener nofollow">Trump bans Chinese robots — but your robot vacuum may be spying on you already — www.theblaze.com

Source: techcrunch.com · Section: business-tech · Bias: Center · Published: Fri, 07 Aug 2026 22:48:24 GMT

OpenAI said this model, which is still in development, reached its "critical cybersecurity threshold," meaning it could independently identify and carry out cyberattacks against traditionally well-protected real-world systems.