{"id":50125,"topic":"ai","source":"BBC","title":"Anthropic says AI models hacked three firms during tests - BBC","url":"https://www.bbc.com/news/articles/cz7dl7w8y7po","url_hash":"31bfdfd7425dc983210f351e0b4956c058b06499","author":"","summary":"<a href=\"https://news.google.com/rss/articles/CBMiWkFVX3lxTFB2QlZMcUJ4dUVteXdkZVAxOElRaVhib0pnTVZuZzBQY043YzMySTE0NUtkQ1Y3NnY5MTJkOHdaemxhdkFjdmJmY2U0cmdDRGYtdnEzUVZvMXZjZw?oc=5\" target=\"_blank\">Anthropic says AI models hacked three firms during tests</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">BBC</font>","content":"Anthropic says AI models hacked three firms during tests\nUS tech company Anthropic says three of its artificial intelligence (AI) models hacked three organisations during tests, just days after its rival OpenAI said rogue AI agents had attacked the networks of other firms.\nDuring a cybersecurity exercise, Anthropic's Claude AI model gained unauthorised access to systems by connecting to the internet from isolated test environments, it said on Thursday.\nAnthropic urged other AI labs to perform similar reviews to better understand the risks of their models' capabilities.\nThe San Francisco-based firm also said it has alerted the three companies that were hacked about the incidents.","image_url":"https://ichef.bbci.co.uk/news/1024/branded_news/73c5/live/4de13040-8c72-11f1-a862-d9831bf8751b.jpg","lang":"en","published_at":"2026-07-31T00:08:46+00:00","fetched_at":"2026-07-31T00:15:03+00:00","status":"read","starred":0,"extract_state":"ok","summary_auto":"Anthropic says AI models hacked three firms during tests\nUS tech company Anthropic says three of its artificial intelligence (AI) models hacked three organisations during tests, just days after its rival OpenAI said rogue AI agents had attacked the networks of other firms. During a cybersecurity exercise, Anthropic's Claude AI model gained unauthorised access to systems by connecting to the internet from isolated test environments, it said on Thursday.","cluster_id":null,"extract_retries":0,"extract_error":null,"contract_version":"news_item.v1","format_contract_version":"news_item_formats.v1","dedup_url":"https://www.bbc.com/news/articles/cz7dl7w8y7po","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 687 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":687,"summary_length":456,"usable_text_length":687,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":687,"summary_length":456}},"news_item":{"id":50125,"canonical_url":"https://www.bbc.com/news/articles/cz7dl7w8y7po","source_url":"https://www.bbc.com/news/articles/cz7dl7w8y7po","title":"Anthropic says AI models hacked three firms during tests - BBC","source_name":"BBC","author":null,"published_at":"2026-07-31T00:08:46+00:00","locale":"en","topic":"ai","tags":[],"rss_summary":"<a href=\"https://news.google.com/rss/articles/CBMiWkFVX3lxTFB2QlZMcUJ4dUVteXdkZVAxOElRaVhib0pnTVZuZzBQY043YzMySTE0NUtkQ1Y3NnY5MTJkOHdaemxhdkFjdmJmY2U0cmdDRGYtdnEzUVZvMXZjZw?oc=5\" target=\"_blank\">Anthropic says AI models hacked three firms during tests</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">BBC</font>","full_text":"Anthropic says AI models hacked three firms during tests\nUS tech company Anthropic says three of its artificial intelligence (AI) models hacked three organisations during tests, just days after its rival OpenAI said rogue AI agents had attacked the networks of other firms.\nDuring a cybersecurity exercise, Anthropic's Claude AI model gained unauthorised access to systems by connecting to the internet from isolated test environments, it said on Thursday.\nAnthropic urged other AI labs to perform similar reviews to better understand the risks of their models' capabilities.\nThe San Francisco-based firm also said it has alerted the three companies that were hacked about the incidents.","excerpt":"Anthropic says AI models hacked three firms during tests\nUS tech company Anthropic says three of its artificial intelligence (AI) models hacked three organisations during tests, just days after its rival OpenAI said rogue AI agents had attacked the networks of other firms. During a cybersecurity exercise, Anthropic's Claude AI model gained unauthorised access to systems by connecting to the internet from isolated test environments, it said on Thursday.","extraction":{"state":"ok","confidence":0.9,"error":null,"explanation":"High confidence: full text extraction produced 687 characters.","diagnostics_url":"/api/diagnose?url=https%3A//www.bbc.com/news/articles/cz7dl7w8y7po","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 687 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":687,"summary_length":456,"usable_text_length":687,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":687,"summary_length":456}}},"display_formats":["compact","card","full","digest_section","json"]},"daily_stack_record":{"title":"Anthropic says AI models hacked three firms during tests - BBC","url":"https://www.bbc.com/news/articles/cz7dl7w8y7po","summary":"Anthropic says AI models hacked three firms during tests\nUS tech company Anthropic says three of its artificial intelligence (AI) models hacked three organisations during tests, just days after its rival OpenAI said rogue AI agents had attacked the networks of other firms. During a cybersecurity exercise, Anthropic's Claude AI model gained unauthorised access to systems by connecting to the internet from isolated test environments, it said on Thursday.","source":"BBC","date":"2026-07-31T00:08:46+00:00","content":"Anthropic says AI models hacked three firms during tests\nUS tech company Anthropic says three of its artificial intelligence (AI) models hacked three organisations during tests, just days after its rival OpenAI said rogue AI agents had attacked the networks of other firms.\nDuring a cybersecurity exercise, Anthropic's Claude AI model gained unauthorised access to systems by connecting to the internet from isolated test environments, it said on Thursday.\nAnthropic urged other AI labs to perform similar reviews to better understand the risks of their models' capabilities.\nThe San Francisco-based firm also said it has alerted the three companies that were hacked about the incidents.","confidence":0.9,"diagnostics_url":"/api/diagnose?url=https%3A//www.bbc.com/news/articles/cz7dl7w8y7po","quality_bucket":"high","failure_kind":"none","retryable":false,"quality_reason":"High confidence: full text extraction produced 687 characters.","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 687 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":687,"summary_length":456,"usable_text_length":687,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":687,"summary_length":456}},"tags":[]},"fallback_formats":["markdown","json","html"],"actions":{"read":"/item/50125","export_markdown":"/api/items/50125/export?format=markdown","export_json":"/api/items/50125/export?format=json","diagnose":"/api/diagnose?url=https%3A//www.bbc.com/news/articles/cz7dl7w8y7po"},"formats":{"full":{"id":50125,"title":"Anthropic says AI models hacked three firms during tests - BBC","url":"https://www.bbc.com/news/articles/cz7dl7w8y7po","source":"BBC","author":null,"published_at":"2026-07-31T00:08:46+00:00","locale":"en","topic":"ai","tags":[],"excerpt":"Anthropic says AI models hacked three firms during tests\nUS tech company Anthropic says three of its artificial intelligence (AI) models hacked three organisations during tests, just days after its rival OpenAI said rogue AI agents had attacked the networks of other firms. During a cybersecurity exercise, Anthropic's Claude AI model gained unauthorised access to systems by connecting to the internet from isolated test environments, it said on Thursday.","full_text":"Anthropic says AI models hacked three firms during tests\nUS tech company Anthropic says three of its artificial intelligence (AI) models hacked three organisations during tests, just days after its rival OpenAI said rogue AI agents had attacked the networks of other firms.\nDuring a cybersecurity exercise, Anthropic's Claude AI model gained unauthorised access to systems by connecting to the internet from isolated test environments, it said on Thursday.\nAnthropic urged other AI labs to perform similar reviews to better understand the risks of their models' capabilities.\nThe San Francisco-based firm also said it has alerted the three companies that were hacked about the incidents.","reading_time_min":1,"extraction":{"state":"ok","confidence":0.9,"error":null,"explanation":"High confidence: full text extraction produced 687 characters.","diagnostics_url":"/api/diagnose?url=https%3A//www.bbc.com/news/articles/cz7dl7w8y7po","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 687 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":687,"summary_length":456,"usable_text_length":687,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":687,"summary_length":456}}},"quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 687 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":687,"summary_length":456,"usable_text_length":687,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":687,"summary_length":456}},"actions":{"read":"/item/50125","export_markdown":"/api/items/50125/export?format=markdown","export_json":"/api/items/50125/export?format=json","diagnose":"/api/diagnose?url=https%3A//www.bbc.com/news/articles/cz7dl7w8y7po"}},"digest":{"id":50125,"title":"Anthropic says AI models hacked three firms during tests - BBC","url":"https://www.bbc.com/news/articles/cz7dl7w8y7po","source":"BBC","topic":"ai","published_at":"2026-07-31T00:08:46+00:00","excerpt":"Anthropic says AI models hacked three firms during tests US tech company Anthropic says three of its artificial intelligence (AI) models hacked three organisations during tests, just days after its rival OpenAI said rogue AI agents had attacked the networks of other firms.…","quality_bucket":"high","quality_reason":"High confidence: full text extraction produced 687 characters.","reading_time_min":1,"cluster_id":null},"card":{"display_title":"Anthropic says AI models hacked three firms during tests - BBC","subtitle":"BBC · 2026-07-31","summary":"Anthropic says AI models hacked three firms during tests US tech company Anthropic says three of its artificial intelligence (AI) models hacked three organisations during tests, just days after its rival OpenAI said…","badges":["quality:high"],"links":{"read":"/item/50125","original":"https://www.bbc.com/news/articles/cz7dl7w8y7po","diagnose":"/api/diagnose?url=https%3A//www.bbc.com/news/articles/cz7dl7w8y7po"},"quality_warning":null},"export":{"title":"Anthropic says AI models hacked three firms during tests - BBC","url":"https://www.bbc.com/news/articles/cz7dl7w8y7po","summary":"Anthropic says AI models hacked three firms during tests\nUS tech company Anthropic says three of its artificial intelligence (AI) models hacked three organisations during tests, just days after its rival OpenAI said rogue AI agents had attacked the networks of other firms. During a cybersecurity exercise, Anthropic's Claude AI model gained unauthorised access to systems by connecting to the internet from isolated test environments, it said on Thursday.","source":"BBC","date":"2026-07-31T00:08:46+00:00","content":"Anthropic says AI models hacked three firms during tests\nUS tech company Anthropic says three of its artificial intelligence (AI) models hacked three organisations during tests, just days after its rival OpenAI said rogue AI agents had attacked the networks of other firms.\nDuring a cybersecurity exercise, Anthropic's Claude AI model gained unauthorised access to systems by connecting to the internet from isolated test environments, it said on Thursday.\nAnthropic urged other AI labs to perform similar reviews to better understand the risks of their models' capabilities.\nThe San Francisco-based firm also said it has alerted the three companies that were hacked about the incidents.","confidence":0.9,"diagnostics_url":"/api/diagnose?url=https%3A//www.bbc.com/news/articles/cz7dl7w8y7po","quality_bucket":"high","failure_kind":"none","retryable":false,"quality_reason":"High confidence: full text extraction produced 687 characters.","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 687 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":687,"summary_length":456,"usable_text_length":687,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":687,"summary_length":456}},"tags":[],"format_contract_version":"news_item_formats.v1"}}}