{"input_url":"https://www.aljazeera.com/news/2026/7/29/how-are-ai-models-able-to-autonomously-hack-others","gnews_decode":"https://www.aljazeera.com/news/2026/7/29/how-are-ai-models-able-to-autonomously-hack-others","http_status":200,"final_url":"https://www.aljazeera.com/news/2026/7/29/how-are-ai-models-able-to-autonomously-hack-others","html_length":205307,"trafilatura_result":"ok","text_length":8448,"text_preview":"Last week, two of OpenAI’s most advanced AI models were reported to have “escaped” a controlled testing environment and hacked Hugging Face, a totally separate AI company, moving from one computer system to another to complete their task.\nReuters reported that the models exploited vulnerable code wr","verdict":"OK"}