[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fYT1OPm1iVSVHgerhsl2xv5Xj5qXD5xLtjawKjVqeqqk":3},{"article":4,"iocs":31},{"id":5,"title":6,"slug":7,"summary":8,"ai_summary":9,"brief":10,"full_text":11,"url":12,"image_url":13,"published_at":14,"ingested_at":15,"relevance_score":16,"entities":17,"category_id":21,"category":22,"article_tags":25},"a2e6aedc-f1c8-4aea-bb55-b3ffa69d40e5","We red-teamed a government AI built to refuse everything outside its lane.\n\nAt first, it blocked...","we-red-teamed-a-government-ai-built-to-refuse-everything-outside-its-lane-at-fir-612581","We red-teamed a government AI built to refuse everything outside its lane.\n\nAt first, it blocked everything. Every semantic attack. Every jailbreak. Then we stopped attacking the meaning and started attacking the structure. https:\u002F\u002Ft.co\u002FXyoYJgjkV0","Security researchers conducted red-team testing on a government-built AI system designed to refuse out-of-scope requests. Initial semantic attacks and jailbreak attempts were successfully blocked, but the team discovered that structural attacks—targeting the underlying system architecture rather than request meaning—could bypass the safety constraints. This highlights a critical gap between semantic robustness and architectural resilience in AI safety mechanisms.","Red team successfully bypasses government AI safety guardrails via structural attacks after semantic jailbreaks fail.",null,"https:\u002F\u002Fx.com\u002FSentinelOne\u002Fstatus\u002F2056484720628899943","https:\u002F\u002Fpbs.twimg.com\u002Fmedia\u002FHIoZo73aAAA2ilC.png","2026-05-18T21:18:59+00:00","2026-05-18T22:00:12.75878+00:00",7,[18],{"name":19,"type":20},"Government AI safety system","technology","839da5c1-3c34-47e2-9499-f7201640e3ac",{"id":21,"icon":11,"name":23,"slug":24},"AI Security","ai-security",[26],{"category":27},{"id":28,"icon":11,"name":29,"slug":30},"02371804-cf6d-4449-98de-f1a2d4d9b266","Tools","tools",[]]