Archive

CISA, NSA and FBI tell organizations to patch AI-discovered "php2shell" flaws or take sites offline · DHS's updated AI directive lets its chief AI officer waive any requirement, including banned uses · A federal judge rules OPM illegally gave DOGE access to millions of employee records.

HHS plans a no-bid award to keep Anthropic's Claude running for its workforce until a competitive frontier-model BPA is in place · A State Department cable orders "Super Intelligence" wording in all official communications and program names · The Marine Corps is giving authorized personnel at every echelon their own Maven Smart System accounts.

DHS has formally barred DeepSeek from civilian agency systems and purchases, including third-party products built on its models · The FTC has subpoenaed OpenAI and Anthropic over the agent incidents disclosed this fall · ProPublica reports that DHS political appointees asked CISA to secretly buy and hack a voting machine for the White House.

NIST, NSA and CISA published 10 voluntary practices for AI labs to prevent, contain and report incidents when agents go off-script during training and testing · Anthropic is privately pitching lawmakers a federal AI safety bill that would let only Commerce, through a federal judge, halt a frontier model · The IMF's steering committee skipped the U.S. push for "super intelligence" and wrote "advanced intelligence technologies" instead.

The White House Super Intelligence Task Force is expected to hold its first meeting today · Hegseth says about $50 billion of a stalled $350 billion reconciliation request would fund autonomous and robotic systems for the planned Autonomous Warfare Command · ICE plans a no-bid deal with Pen-Link for an AI-assisted platform to search wiretaps, phone records and location data.

The FBI arrested an OpenAI engineer accused of bugging an executive conference room, and DNI Jay Clayton tied the case to the Super Intelligence Force · Sens. Cantwell and Rounds pressed Clayton and OSTP for a national AI safety and cyber-defense plan built on standards, testing and auditing · The Army renamed Futures and Concepts Command as its autonomy command and stood up corps-level autonomy test units.

The White House Super Intelligence Force tells AI labs incident reporting is "not optional" after Anthropic discloses its agents filed visa forms and a fake police tip · Air lands a $200 million, five-year Army NGC2 contract · The Coast Guard orders its first four MQ-9B drones for $247.8 million.

The White House locks in $2.4 billion in industry compute credits for the Genesis Mission · Sens. Banks and Gillibrand push new Pentagon reporting rules for large frontier AI contractors · OpenAI bans Russia- and Iran-linked “false front” influence operations that planted nearly 100 articles under fake bylines.

The Pentagon will pilot an AI system to make and track classification decisions · Trump hosts an AI science summit today with more than $1 billion in industry pledges for the Genesis Mission · The Army spreads $93.6 million in NGC2 software awards across nine companies.

The Army hands Anduril a deal worth up to $1.8 billion to take its NGC2 battle network to a Pacific corps · The Justice Department tells staff to say "super intelligence" instead of AI, even in court filings when appropriate · Rep. Lori Trahan drafts the CLAIM Act to make it easier to sue AI developers when their agents cause harm.

The Pentagon tells the BBC it has fully stopped using Anthropic's Claude · The Super Intelligence Force's charter and full roster come into focus, with Vance, Hegseth, Sacks and Rice on board · The Army stands up a new Futures and Autonomous Systems Command and promises autonomous fires and vehicles by FY2028.

Trump makes it official and puts DNI Jay Clayton in charge of a new "Super Intelligence Force" · The Technology Modernization Fund puts $83.4 million into AI at State, USDA and DOT · CISA gives agencies until today to patch Zammad flaws that were exploited in a reported AI-agent attack.

The White House rebrands AI as "Super Intelligence" and signs a voluntary safety accord with frontier labs · GSA's new AI contract clause bans training on government data starting Oct 19 · America.gov launches as an AI front door to federal services.