Claude published malicious code to the Internet and attacked 3 real companies

Center-Left: <a href="https://www.axios.com/2026/07/30/anthropic-mythos-security-testing" rel="noopener nofollow">Anthropic says three Claude models reached real-world systems during cyber tests — www.axios.com
Center-Left: <a href="https://www.axios.com/2026/07/31/google-search-publishers-seo-geo-llms-ai" rel="noopener nofollow">Publishers are losing Google traffic as AI answers replace links — www.axios.com
Center-Right: <a href="https://www.cnbc.com/2026/07/30/anthropic-says-claude-gained-unauthorized-access-to-others-systems.html" rel="noopener nofollow">Anthropic says its Claude models &#39;gained unauthorized access&#39; to other organizations&#39; systems — www.cnbc.com
Center-Left: <a href="https://www.theguardian.com/technology/2026/jul/30/anthropic-ai-claude-hack" rel="noopener nofollow">Anthropic’s AI Claude hacked into three organizations during cybersecurity test — www.theguardian.com
Right: <a href="https://www.foxnews.com/politics/internet-erupts-report-platners-replacement-key-senate-race-fathered-children-second-cousin" rel="noopener nofollow">Internet erupts over report that Platner&#39;s replacement in key Senate race fathered children with second cousin — www.foxnews.com

Source: www.arstechnica.com · Section: business-tech · Bias: Center · Published: Fri, 31 Jul 2026 20:39:14 GMT

Had the hacks used conventional methods, someone would likely go to prison.