18 stories
Federal AI & Tech Digest: Monday, Oct 19, 2026
No saved stories in this issue.
White House & Executive Action
IMF members call for more support amid war shocks, 'advanced intelligence' risks
(2 minute read) Update: Friday's digest reported Treasury pressing IMF members to say "super intelligence" in the International Monetary and Financial Committee statement. The final statement, released Oct. 17 by committee chair and Saudi Finance Minister Mohammed Al-Jadaan, used neither "artificial intelligence" nor "super intelligence." It said "advanced intelligence technologies" have boosted investment and trade but that their uncertain economic impact, together with rising bond yields, "could amplify macro-financial vulnerabilities." Al-Jadaan said all members agreed to the language. It's a sign that other governments aren't yet willing to adopt the administration's SI terminology in international forums.
US's AI task force sets path for study and reviews, leader tells CNBC
(1 minute read) Update: the day after the Super Intelligence Force's first meeting, chair and DNI Jay Clayton told CNBC the group set up three committees: reporting and response under Emil Michael, critical infrastructure under Scott Kupor, and law enforcement and responsibility under Andrew Ferguson. He said David Sacks will lead the risk study, that workforce and economic impacts will be reported to the Labor Department, and that any AI wrongdoing would be enforced through the FTC or the Justice Department. That means enforcement will rely on existing consumer-protection and criminal law rather than a new regulator.
Congress & Legislation
Anthropic's proposed AI safety bill only kicks in if 50 deaths or $1 billion in damages are on the line
(3 minute read) Citing Politico, Tom's Hardware reports that Anthropic is quietly circulating draft legislation under which an independent evaluator registered with the Commerce Department would periodically assess frontier labs and their models. The evaluator would have 72 hours to report an imminent risk of a bioweapon, an AI-enabled cyberattack or loss of control over a model or agents that could cause more than $1 billion in damage or more than 50 deaths or serious injuries. Commerce would then take the evidence to a federal judge in D.C., who would decide whether to halt the release, and no other agency could order a lab to stop. Anthropic says the draft reflects the Advanced AI Framework it has been pushing since June. With the SI Force leaning on existing law, this gives Congress an industry-written model for pre-release intervention that also narrows the government's power to step in.
Tech bills of the week: Cutting federal funding for Flock; Making data centers shoulder their power costs; and more
(3 minute read) Rep. Rashida Tlaib's Get the Flock Out of Our Communities Act would bar federal funds from buying or operating automated mass surveillance systems, including Flock's AI-powered cameras and facial recognition. Rep. Eugene Vindman's bill would require data centers drawing at least 20 megawatts to pay the full incremental cost of their electricity. Rep. Yvette Clarke reintroduced a bill requiring political ads that use AI-generated images or video to say so. Combined with Sen. Moreno's RECLAIM Act from last week, data center costs and AI surveillance are becoming campaign-season legislative themes. All of these bills were introduced last week.
Acquisition & Contracts
DISA faces another protest over Enclave contract expansion
(3 minute read) Competitive Range Solutions filed a GAO protest on Oct. 8 challenging DISA's plan to add the migration of combatant commands to DoDNet onto Leidos' $11 billion Defense Enclave Services contract without competition. It follows a GDIT protest that was dismissed after DISA promised new market research. DISA then gave industry several days to answer one sources sought notice and 24 hours to answer a second, before concluding again that Leidos is the only capable source for the department's Sept. 30, 2028, deadline. The protest tests how far DOD can stretch existing vehicles in the name of speed, and DISA's justification cites network security "exacerbated by AI."
Core4ce challenges removal from $1.4B RETS contract competition
(1 minute read) Core4ce has protested to GAO after DOD removed it from the competition for the $1.4 billion Research Development Test Evaluation Engineering and Technical Support (RETS) multiple-award contract. The company argues DOD didn't investigate a possible organizational conflict of interest involving a competitor allegedly affiliated with a firm that helped prepare the solicitation and review bids. GAO's decision is due by Jan. 19, and DOD can't make awards until then. RETS consolidates mission engineering, prototyping, and modeling and simulation work now held by incumbents including Modern Technology Solutions, American Systems and Applied Research Associates.
Marine Corps Cyberspace Environment (MCCE) Operational Support Services (OSS) and Enterprise AI Integration
(2 minute read) An Oct. 16 amendment to this Department of the Navy sources sought notice for Quantico moved the response deadline from Oct. 19 to 1 p.m. on Nov. 3. Questions and answers are coming in a fourth amendment. The Marine Corps says the core requirement is skilled IT touch-labor (NAICS 541513), and any AI or automation tooling must be built into the service delivery model to augment the workforce. It also now wants industry feedback on buying "hours of service" or service outcomes rather than individual labor categories. Bidders should present AI as a productivity layer on staff services rather than as a standalone product.
DIU Opens SWAP-USV Sprint 2 for Maritime Counter-Air Systems
(4 minute read) DIU and the Direct Reporting Portfolio Manager for Unmanned Systems opened the second sprint of the SWAP-USV prize challenge. It is looking for production-ready unmanned surface vessels that combine the vessel, launcher and weapon to defeat airborne threats, with a range of at least 200 nautical miles and the ability to deploy forward within 120 days. Sprint 2 offers up to $40 million in prizes, and $200 million is budgeted for follow-on prototype or production OTAs. Pitch decks of up to 15 slides are due Nov. 29, with a three-week government evaluation at Little Creek, Va., in jammed environments and in-water testing in early 2027. Autonomy firms get a direct path to production money if their systems pass.
Army MICC Records $6.8B in FY 2026 Contracting Actions
(3 minute read) The Army's Mission and Installation Contracting Command recorded about $6.8 billion across more than 21,450 contracting actions in FY2026, including $2.9 billion to small businesses, plus more than $520 million in purchase card buys. That is an increase of more than $1 billion over FY2025, which MICC attributes to facility modernization work it took on from the Army Corps of Engineers, broader use of OTAs, SBIR Phase III and MATOCs, and a new contracting brigade. It hit its small business goals for the 12th straight year while operating with 30% fewer personnel. This is one more sign that Army buying is shifting toward faster, nontraditional authorities.
Defense & National Security
US Navy deploys unmanned maritime drone to support carrier strike group in historic first
(3 minute read) Acting Navy Secretary Hung Cao said the Theodore Roosevelt Carrier Strike Group deployed on Oct. 8 with a medium-size unmanned surface vessel, the first carrier strike group ever to deploy with one. The group is headed toward the Middle East. The Navy hasn't said which USV it is, and Leidos, which built the Sea Hunter and Seahawk prototypes, referred questions to the Navy. The deployment moves the Navy's "Golden Fleet" hybrid concept from experimentation into a real-world operation.
Army feels 'growing pains' as it integrates Anduril NGC2 tech with older comms
(6 minute read) In September's Lightning Surge 5 exercise in Hawaii, the 25th Infantry Division's first attempt to run Anduril's Lattice NGC2 software over its existing tactical radios "wasn't seamless," according to Brig. Gen. Shane Taylor. Lockheed Martin, the integrator, says moving data through the system "just is not at the state it should be." Anduril pointed to mismatched radio waveforms and configurations and a mid-exercise request to run Lattice on tablets. The 25th ID's mix of legacy gear looks more like the rest of I Corps than the 4th ID's fresh fielding, so these fixes set the baseline for wider fielding this fiscal year, which the continuing resolution is already slowing.
NSA Issues Zero Trust Guidance for Defense Operational Technology
(2 minute read) NSA's new Cybersecurity Information Sheet, "Hardening Department of War Critical Operational Technology Environments with Zero Trust Principles," is aimed at owners and operators of National Security Systems, Department of War and Defense Industrial Base networks. Federal News Network reports it lists 84 activities, 26 of them high-impact steps organizations can take now, to limit unauthorized access, lateral movement and persistence in legacy OT. NSA named energy, water, agriculture, public health and government facilities as sectors at risk and set no implementation deadline. It gives OT vendors and integrators a concrete checklist ahead of the September 2027 zero trust deadline.
Civilian Agency Initiatives
Top Management and Performance Challenges Facing the Department of Commerce in Fiscal Year 2027
(30 minute read) Commerce's inspector general warns that the Bureau of Industry and Security lacks the capacity to police exports under the American AI Exports Program. BIS had 17 export control officers worldwide as of September 2026, with none resident in Africa or South America, and expects 25 by year's end. The IG also found its end-use check procedures poorly documented and its enforcement systems outdated, which raises the risk that advanced chips and models are diverted to adversaries. The report also flags NIST's National Vulnerability Database, where submissions rose 263% from 2020 to 2025, as likely to fall further behind as SI tools spread. As the administration pushes the "American AI stack" abroad, the enforcement side hasn't kept pace.
Task force driving Coast Guard's zero trust journey for OT, IT systems
(5 minute read) The Coast Guard set up a dedicated zero trust task force to meet the DOD and OMB September 2027 deadline. Capt. Patrick Thompson said the service is "a little bit behind" and has to move 300 applications to modern identity management. The task force is a logistics cell that coordinates router updates and other work across 300 ships and OT systems at more than 1,500 locations, and it is deciding which apps to modernize, merge or retire. The service recently modernized one app with low-code tools in three weeks instead of the expected one to two years. Expect identity, SD-WAN and low-code work to be the near-term buys.
GAO Targets AI, Cybersecurity Oversight in New Strategic Plan
(2 minute read) GAO's FY2027–2032 strategic plan, released Oct. 16, puts AI and cybersecurity at the center of its oversight. It plans to review federal AI deployment and its risks, the effectiveness of AI policy and regulatory frameworks, IT acquisition and software practices, and national security risks from adversarial AI, autonomous systems and AI-enabled decision-making. GAO will also expand its own use of AI and analytics, test tools in its Innovation Lab, and streamline hiring for AI, data science and cyber roles. Agencies should expect more audits like last week's report on the Pentagon's responsible AI council.
AI Safety, Security & Incidents
Key Practices for Incident Readiness and Response in Super Intelligence Training and Evaluation
(2 minute read) On Oct. 16, NIST's Center for Advancing Innovation and Standards for Super Intelligence (CAISSI), NSA's Super Intelligence Security Center and CISA published 10 voluntary practices for incidents in which AI agents take "significant uninstructed actions" that break operator controls or human oversight during training and evaluation. The practices cover five stages: preparation, detection, response, recovery and reporting. They include least-privilege access and strict network allowlists, monitors the agents can't tamper with, kill mechanisms, tamper-resistant logs, revoking credentials an agent accessed, notifying affected parties, voluntarily reporting incidents and near misses to CISA, and public disclosure. The guidance builds on NIST CSF 2.0 and SP 800-61r3 and draws on CAISSI's own evaluation work. It is the first federal playbook for the kinds of agent incidents labs have disclosed this fall, and it could inform the lab reporting protocol the SI Force is drafting.
G7 to launch exercises to guard against AI cyber attacks
(2 minute read) G7 finance ministers and central bank governors, the U.S. among them, agreed in Bangkok to run tabletop exercises simulating AI-driven cyberattacks on financial institutions, Japanese Finance Minister Satsuki Katayama said. The results will test existing protocols and be used to revise incident manuals, and private banks may take part depending on the scenario. The deal came after the IMF warned that an AI-enabled attack on a critical financial service could undermine confidence. Treasury and financial regulators are now planning for frontier-AI cyber risk alongside the SI Force's critical infrastructure work.
Industry & Vendor News
⚠️ Accenture Federal Services and Google Public Sector Establish Gemini Launchpads to Power Purpose-Built Agents Across Federal Missions
(4 minute read) Accenture Federal Services and Google Public Sector announced "Gemini launchpads," a hub of forward-deployed engineers who embed with agencies to build and scale Gemini agents connected to their systems of record. They are also developing pre-built agents for HR, finance and supply chain. Accenture is training an initial cohort of Google Cloud-skilled engineers as part of a planned 1,000-person forward-deployed engineering workforce, and it offers agent-building workshops at The Forge in Washington. Embedding engineers inside agencies is becoming a common way for integrators to compete to move agencies from agent pilots into production. This is a company announcement and names no agency customers.
Upcoming Deadlines
- Oct 19 (today): Comments due on the FAR overhaul second tranche, including data-rights FAR Case 2026-011; GSA's LLM data-safeguarding clause (GSAR 552.239-7001) takes effect (Federal News Network, FedScoop)
- Oct 19 (today): FAA answers due to House Democrats on SMART/FMDS pre-deployment testing, training and oversight (Federal News Network) (new)
- Oct 19 (today): FCC reply comments due on the Club for Growth AI political robocall waiver, docket 02-278 (Communications Daily)
- Oct 19 (today): FDA Center for Tobacco Products sources sought responses due on data management and AI prototypes (FedReady)
- Army Project Griffin: two cyber-defense agent awards were expected Oct 16, but no public announcement had turned up as of this morning (Breaking Defense) (updated)
- Oct 21: CBP controlled UAS platform (SDVOSB set-aside) proposals due (ExecutiveGov)
- Oct 21, 5:00 p.m. ET: AFRL JADPACT II sources sought responses due (Govly)
- Oct 23, 2:00 p.m. ET: ICE OCIO enterprise AI/data/tech strategy and Agentic Software Factory RFI (STELLA) responses due (SAM.gov)
- Oct 23: U.S. Tech Force job announcement closes (Federal News Network)
- Oct 23: Army Project Gatekeeper software-vetting sandbox goes live (Breaking Defense)
- Oct 23: Responses due on SDA PEREGRINE Sentry sources sought (ExecutiveGov)
- Oct 26, 9:00 a.m. ET: ICE RFI responses due on an SI-powered detention communications and video intelligence suite (SAM.gov)
- Oct 26, 10:00 a.m. ET: Capability responses due on ICE's planned sole-source Pen-Link platform (Project Salt Box)
- Oct 26: RSVP deadline for DOD's Next Gen IAC MAC industry day (Washington Technology)
- Oct 28, 11:59 p.m. ET: Comments due on NIH's proposed Office of the Chief AI Officer reorganization (Federal Register)
- Oct 28: DHA Data Exchange Platform CSO Phase I proposals due (ExecutiveGov)
- Oct 28: Registration closes for ARPA-H's Nov 6 SURPASS Proposers' Day (ARPA-H)
- Oct 29: CBP Border Patrol Technical Operations Center (BPTOC) RFI responses due (ExecutiveGov)
- ~Oct 30: DRPM-UxS plan of action and milestones for the Autonomous Warfare Command due to Hegseth (GovConWire)
- Nov 3, 1:00 p.m.: Marine Corps MCCE Operational Support Services and Enterprise AI Integration sources sought responses due, extended from Oct 19 (Oppward) (new)
- Nov 4, 11 a.m. ET: CMS fraud-prevention contractor operating model RFI responses due (Nextgov/FCW)
- Nov 6, 6 p.m. ET: Registration closes for Army Joint Innovation Outpost Nov 17 virtual industry day (Intelligence Community News)
- Nov 12: GSA deadline for AbilityOne Commission country-of-origin verification plan (Nextgov/FCW)
- Nov 14: Applications due for the Marine Corps "Mindgame" human-vs-AI wargaming tournament (DefenseScoop)
- Mid-November: Army targets Login.gov integration to open Project Gatekeeper to commercial vendors (Breaking Defense)
- ~Nov 28: White House science adviser's proposed legislative definition of "Super Intelligence" due, ~60 days after EO 14434 (EO 14434)
- Nov 29: DIU SWAP-USV Sprint 2 pitch decks due (ExecutiveGov) (new)
- Nov 30: ARPA-H SURPASS solution summaries due; CBP non-intrusive inspection logistics (OASIS+ 8(a)) solicitation expected (ARPA-H, ExecutiveGov)
- Dec 1: CBP Maritime Approaches Surveillance Systems (MAST) solicitation expected, $100M+ (ExecutiveGov)
- December: JIATF-401 directed energy "shoot off" (Breaking Defense)
- Jan 1, 2027: Service secretaries' updated cyber incentive pay plans due under the C-MIP transition (DefenseScoop)
- Jan 19, 2027: GAO decision due on Core4ce's RETS protest; no RETS awards until then (Washington Technology) (new)
- ~Feb 12, 2027: End of the Super Intelligence Force's 120-day window (Semafor)
- ~April 2027: Pentagon ACME automated classification pilot initial deployment (DefenseScoop)
- September 2027: DOD/OMB zero trust deadline the Coast Guard is racing to meet (Federal News Network) (new)
- Sept 30, 2027: C-MIP legacy-pay transition period ends (DefenseScoop)
- Oct 1, 2027: Target activation date for the four-star Autonomous Warfare Command (MeriTalk)
- Sept 30, 2028: DOD deadline to migrate combatant commands to DoDNet, the work at issue in the DISA Enclave protests (Federal News Network) (new)
Upcoming Events
- AFCEA Mission Advantage: From Data Ownership to Operational AI, Oct 20, 1:00 p.m., Huntsville.
- Google Public Sector Summit, Oct 20, Washington, D.C.
- CISA Strategic Cyber Tools Buy virtual industry day, Oct 20, 9 a.m. ET, Google Meet. Registration has closed.
- FedInsider Innovation Summit 2026: AI Dominates the Landscape, Oct 21, Washington, D.C.
- ACT-IAC Federal Insights Exchange with USDA's Tony Brannum, Oct 22. USDA cybersecurity and preparing for AI's security implications.
- AFCEA TechNet Indo-Pacific, Oct 26–29, Hilton Hawaiian Village, Honolulu.
- Red Hat Government Symposium 2026, Oct 28, Ronald Reagan Building, Washington, D.C.
- ICE ORBIT industry day, Oct 29, Arlington, Va., and virtual.
- Potomac Officers Club 2026 FedCiv Summit, Oct 29.
- DOD Next Gen IAC MAC industry day, Oct 30. RSVP by Oct 26.
- ACT-IAC Federal Insights Exchange: Operationalizing AI for Enterprise Transformation, Nov 5, in person. How agencies move AI from experiments into mission workflows at enterprise scale. (new)
- GovCIO Media & Research AI Summit 2026, Nov 6, Hyatt Regency Tysons Corner, Tysons, Va. Sessions on the AI Action Plan, the Genesis Mission, agentic AI and governance, with speakers including NIST's senior SI adviser. (new)
- ACT-IAC Workforce Innovation Summit, Nov 6, in person. Federal technology and workforce transformation. (new)
- ACT-IAC Service Delivery and Experience Summit, Nov 13, in person. How AI is reshaping digital service delivery. (new)
- Public Service AI Summit 2026, Nov 17, Washington, D.C. (in person by invitation) and virtual. Hosted by the Partnership for Public Service's AI Center for Government, it launches the first Public Service AI Excellence Awards. (new)
- AFCEA DC November Luncheon: Decision Dominance through Awareness: OSINT and Strategic Intelligence, Nov 19, Army Navy Country Club, Arlington, Va. Agentic AI for real-time open-source intelligence and how to trust results when adversaries seed deception. (new)
Reliability Score: 5 official source · 4 established outlet · 3 analysis, vendor or third-party copy · 2 needs a second source. Ask me about any story for more detail.