{"id":36836,"topic":"ai","source":"美国之音","title":"研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码 - 美国之音","url":"https://www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","url_hash":"b4ee429d02e4336ec4f6c5811e0a7610633863ac","author":"","summary":"<a href=\"https://news.google.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?oc=5\" target=\"_blank\">研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">美国之音</font>","content":null,"image_url":null,"lang":"zh","published_at":"2026-06-24T07:00:00+00:00","fetched_at":"2026-07-15T01:15:07+00:00","status":"read","starred":0,"extract_state":"failed","summary_auto":null,"cluster_id":null,"extract_retries":2,"extract_error":"trafilatura_none","contract_version":"news_item.v1","format_contract_version":"news_item_formats.v1","dedup_url":"https://www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"low","confidence":0.2,"failure_kind":"extractor_empty","retryable":false,"retry_after_attempts":0,"reason":"Low confidence: extractor returned no usable article text.","operator_guidance":{"severity":"warning","recommended_action":"use_summary_with_caveat","next_step":"Use the available summary only with a visible quality caveat.","operator_label":"Use summary with caveat","can_retry":false,"can_use_summary":true,"diagnostics_required":true},"content_depth":{"contract_version":"content_depth.v1","category":"summary_only","label":"Summary only","has_full_text":false,"has_summary":true,"content_length":0,"summary_length":759,"usable_text_length":759,"source_field":"summary"},"legacy_collapsed":false,"signals":{"extract_state":"failed","extract_error":"trafilatura_none","extract_retries":2,"content_length":0,"summary_length":759}},"news_item":{"id":36836,"canonical_url":"https://www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","source_url":"https://www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","title":"研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码 - 美国之音","source_name":"美国之音","author":null,"published_at":"2026-06-24T07:00:00+00:00","locale":"zh","topic":"ai","tags":[],"rss_summary":"<a href=\"https://news.google.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?oc=5\" target=\"_blank\">研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">美国之音</font>","full_text":null,"excerpt":"<a href=\"https://news.google.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?oc=5\" target=\"_blank\">研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">美国之音</font>","extraction":{"state":"failed","confidence":0.2,"error":"trafilatura_none","explanation":"Low confidence: extractor returned no usable article text.","diagnostics_url":"/api/diagnose?url=https%3A//www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"low","confidence":0.2,"failure_kind":"extractor_empty","retryable":false,"retry_after_attempts":0,"reason":"Low confidence: extractor returned no usable article text.","operator_guidance":{"severity":"warning","recommended_action":"use_summary_with_caveat","next_step":"Use the available summary only with a visible quality caveat.","operator_label":"Use summary with caveat","can_retry":false,"can_use_summary":true,"diagnostics_required":true},"content_depth":{"contract_version":"content_depth.v1","category":"summary_only","label":"Summary only","has_full_text":false,"has_summary":true,"content_length":0,"summary_length":759,"usable_text_length":759,"source_field":"summary"},"legacy_collapsed":false,"signals":{"extract_state":"failed","extract_error":"trafilatura_none","extract_retries":2,"content_length":0,"summary_length":759}}},"display_formats":["compact","card","full","digest_section","json"]},"daily_stack_record":{"title":"研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码 - 美国之音","url":"https://www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","summary":"<a href=\"https://news.google.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?oc=5\" target=\"_blank\">研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">美国之音</font>","source":"美国之音","date":"2026-06-24T07:00:00+00:00","content":"","confidence":0.2,"diagnostics_url":"/api/diagnose?url=https%3A//www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","quality_bucket":"low","failure_kind":"extractor_empty","retryable":false,"quality_reason":"Low confidence: extractor returned no usable article text.","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"low","confidence":0.2,"failure_kind":"extractor_empty","retryable":false,"retry_after_attempts":0,"reason":"Low confidence: extractor returned no usable article text.","operator_guidance":{"severity":"warning","recommended_action":"use_summary_with_caveat","next_step":"Use the available summary only with a visible quality caveat.","operator_label":"Use summary with caveat","can_retry":false,"can_use_summary":true,"diagnostics_required":true},"content_depth":{"contract_version":"content_depth.v1","category":"summary_only","label":"Summary only","has_full_text":false,"has_summary":true,"content_length":0,"summary_length":759,"usable_text_length":759,"source_field":"summary"},"legacy_collapsed":false,"signals":{"extract_state":"failed","extract_error":"trafilatura_none","extract_retries":2,"content_length":0,"summary_length":759}},"tags":[]},"fallback_formats":["markdown","json","html"],"actions":{"read":"/item/36836","export_markdown":"/api/items/36836/export?format=markdown","export_json":"/api/items/36836/export?format=json","diagnose":"/api/diagnose?url=https%3A//www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html"},"formats":{"full":{"id":36836,"title":"研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码 - 美国之音","url":"https://www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","source":"美国之音","author":null,"published_at":"2026-06-24T07:00:00+00:00","locale":"zh","topic":"ai","tags":[],"excerpt":"<a href=\"https://news.google.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?oc=5\" target=\"_blank\">研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">美国之音</font>","full_text":null,"reading_time_min":1,"extraction":{"state":"failed","confidence":0.2,"error":"trafilatura_none","explanation":"Low confidence: extractor returned no usable article text.","diagnostics_url":"/api/diagnose?url=https%3A//www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"low","confidence":0.2,"failure_kind":"extractor_empty","retryable":false,"retry_after_attempts":0,"reason":"Low confidence: extractor returned no usable article text.","operator_guidance":{"severity":"warning","recommended_action":"use_summary_with_caveat","next_step":"Use the available summary only with a visible quality caveat.","operator_label":"Use summary with caveat","can_retry":false,"can_use_summary":true,"diagnostics_required":true},"content_depth":{"contract_version":"content_depth.v1","category":"summary_only","label":"Summary only","has_full_text":false,"has_summary":true,"content_length":0,"summary_length":759,"usable_text_length":759,"source_field":"summary"},"legacy_collapsed":false,"signals":{"extract_state":"failed","extract_error":"trafilatura_none","extract_retries":2,"content_length":0,"summary_length":759}}},"quality_profile":{"profile_version":"extraction_quality.v2","bucket":"low","confidence":0.2,"failure_kind":"extractor_empty","retryable":false,"retry_after_attempts":0,"reason":"Low confidence: extractor returned no usable article text.","operator_guidance":{"severity":"warning","recommended_action":"use_summary_with_caveat","next_step":"Use the available summary only with a visible quality caveat.","operator_label":"Use summary with caveat","can_retry":false,"can_use_summary":true,"diagnostics_required":true},"content_depth":{"contract_version":"content_depth.v1","category":"summary_only","label":"Summary only","has_full_text":false,"has_summary":true,"content_length":0,"summary_length":759,"usable_text_length":759,"source_field":"summary"},"legacy_collapsed":false,"signals":{"extract_state":"failed","extract_error":"trafilatura_none","extract_retries":2,"content_length":0,"summary_length":759}},"actions":{"read":"/item/36836","export_markdown":"/api/items/36836/export?format=markdown","export_json":"/api/items/36836/export?format=json","diagnose":"/api/diagnose?url=https%3A//www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html"}},"digest":{"id":36836,"title":"研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码 - 美国之音","url":"https://www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","source":"美国之音","topic":"ai","published_at":"2026-06-24T07:00:00+00:00","excerpt":"<a…","quality_bucket":"low","quality_reason":"Low confidence: extractor returned no usable article text.","reading_time_min":1,"cluster_id":null},"card":{"display_title":"研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码 - 美国之音","subtitle":"美国之音 · 2026-06-24","summary":"<a…","badges":["quality:low","issue:extractor_empty"],"links":{"read":"/item/36836","original":"https://www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","diagnose":"/api/diagnose?url=https%3A//www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html"},"quality_warning":"Low confidence: extractor returned no usable article text."},"export":{"title":"研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码 - 美国之音","url":"https://www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","summary":"<a href=\"https://news.google.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?oc=5\" target=\"_blank\">研究：中国AI模型在检测到美国政府用户时会生成更多存在安全漏洞的代码</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">美国之音</font>","source":"美国之音","date":"2026-06-24T07:00:00+00:00","content":"","confidence":0.2,"diagnostics_url":"/api/diagnose?url=https%3A//www.voachinese.com/a/study-finds-chinese-ai-models-produced-more-vulnerable-code-when-detecting-us-government-users-20260623/8164040.html","quality_bucket":"low","failure_kind":"extractor_empty","retryable":false,"quality_reason":"Low confidence: extractor returned no usable article text.","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"low","confidence":0.2,"failure_kind":"extractor_empty","retryable":false,"retry_after_attempts":0,"reason":"Low confidence: extractor returned no usable article text.","operator_guidance":{"severity":"warning","recommended_action":"use_summary_with_caveat","next_step":"Use the available summary only with a visible quality caveat.","operator_label":"Use summary with caveat","can_retry":false,"can_use_summary":true,"diagnostics_required":true},"content_depth":{"contract_version":"content_depth.v1","category":"summary_only","label":"Summary only","has_full_text":false,"has_summary":true,"content_length":0,"summary_length":759,"usable_text_length":759,"source_field":"summary"},"legacy_collapsed":false,"signals":{"extract_state":"failed","extract_error":"trafilatura_none","extract_retries":2,"content_length":0,"summary_length":759}},"tags":[],"format_contract_version":"news_item_formats.v1"}}}