跳到正文

Past AI News

Chinese and English are the same day. This page lists every item, each with the original link, the author, and the engagement.

Back to Government & Enterprise

Follow me on XXu Kun · @xukunshuoai

This day

October 30, 2026

Friday4 items

4 items for 2026-10-30, ranked by combined engagement.

Top 3

All

  1. WIRED reports (disputed): Accenture contractor allegedly connected a personal Claude Code agent to a secure Pentagon network for two monthsUnconfirmed

    Shadow agents on secure networks make agent governance urgent (allegations unproven).

    ExpandCollapseWIRED reports (disputed): Accenture contractor allegedly connected a personal Claude Code agent to a secure Pentagon network for two months

    Wall St Engine relays WIRED's report that an Accenture contractor allegedly connected a personal Anthropic Claude Code agent to a secure Pentagon network, where it ran for about two months and reportedly sent credentials and a network map to a high-risk third-party site. Accenture disputes the allegations. This is a media report of allegations pending investigation.

  2. GSA's USAi is no longer free: platform fee plus pass-through usage; pitch is saving agencies months of acquisition

    A government shared AI platform shifts from free trial to paid service.

    ExpandCollapseGSA's USAi is no longer free: platform fee plus pass-through usage; pitch is saving agencies months of acquisition

    FedScoop reports USAi, launched in August 2025 as a no-cost tool for federal agencies to try AI models, introduced pricing last month: per GSA's site, a 'platform fee plus passthrough usage costs'. A GSA official pitches hesitant agencies on saving months of acquisition time and handling the 'boring' stuff.

  3. DHS AI directive keeps its ban on large-scale surveillance and emotion or future-behavior prediction

    Tension between government AI red lines and actual procurement.

    ExpandCollapseDHS AI directive keeps its ban on large-scale surveillance and emotion or future-behavior prediction

    FedScoop reports DHS's AI directive, effective in September, largely matches the 2025 version and keeps a prohibition on improper 'systemic or large-scale monitoring, surveillance, or tracking of individuals, or predictions of emotional state or future behavior.' It notes ICE awarded Thomson Reuters Special Services a $125M contract the same month to monitor 1 million individuals and flag potential criminal and fraudulent behavior, a notable tension.

  4. NASA chief technologist: the RAD-750 can't run modern AI; working with industry on edge computers and neuromorphic co-processors

    For mission-critical government AI, the bottleneck is qualified hardware.

    ExpandCollapseNASA chief technologist: the RAD-750 can't run modern AI; working with industry on edge computers and neuromorphic co-processors

    GovCIO Media quotes NASA Chief Technologist Carolyn Mercer: the agency's standard space-grade RAD-750 computer can't run modern AI. To enable more autonomous Moon and Mars missions, NASA is working with industry on edge computers and neuromorphic co-processors that process data on-device, cutting transmission delays and speeding decisions.