Pentagon Launches 'AI-First' Strategy Under Hegseth as Lawmakers and Critics Raise Warnings on Military, Economic and Ethical Risks

WASHINGTON (AP) — The U.S. Department of Defense, under new Secretary Pete Hegseth, has prioritized artificial intelligence across its operations, issuing directives in early 2026 to accelerate AI adoption in warfighting, intelligence and enterprise functions. On January 9, 2026, Hegseth reportedly issued a memorandum directing the Pentagon to become an 'AI-first' warfighting institution, according to an article by Janet Abou-Elias and William D. Hartung in The Nation published April 7, 2026. Three days later, on January 12, 2026, the department launched an 'AI Acceleration Strategy,' as detailed in a DoD PDF on media.defense.gov, which outlines three pillars—warfighting, intelligence and enterprise—and introduces seven 'Pace-Setting Projects' (PSPs) aimed at rapid AI integration, such as AI-enabled battlefield decision support and intelligence-to-action systems.

The strategy emphasizes 'wartime speed' in research, development and procurement, framing delays and safeguards as obstacles to maintaining U.S. technological superiority over China, per the DoD document. Critics, including Abou-Elias and Hartung, argue this haste risks flawed systems, excessive costs and ethical lapses, citing historical examples like Vietnam's 'electronic battlefield' and precision-guided munitions in the 1991 Gulf War, though specific claims such as a New York Times quote on Gen. William C. Westmoreland and a Government Accountability Office statistic on munitions tonnage (11 tons guided and 44 tons unguided per target) could not be independently verified in searches of archives. The Pentagon has not issued a public denial of these historical interpretations but maintains that AI represents a 'revolutionary component' for military posture.

On January 26, 2026, the Army awarded Salesforce a 10-year, $5.6 billion contract for AI-enabled systems to consolidate data and enhance decision-making, as reported by Salesforce investor releases and GovCon Wire, building on a decade of prior collaboration. Abou-Elias and Hartung in The Nation portrayed this as increasing private sector influence via venture capital and open-ended contracts, potentially reducing accountability, though DoD officials have described it as boosting 'mission readiness' without commenting on oversight concerns.

Broader AI policy shifts under President Donald Trump include a December 2025 executive order granting the attorney general authority to challenge state AI laws hindering 'U.S. global AI dominance,' potentially withholding federal funds, according to a New York Times report cited in a Nation editorial by Katrina vanden Heuvel and John Nichols on April 7, 2026. In March 2026, the White House released a 'National AI Legislative Framework' emphasizing deregulation and federal preemption, as analyzed by law firms Sidley Austin and Holland & Knight. The framework includes recommendations for child protections, workforce training and IP safeguards, per White House documents, though vanden Heuvel and Nichols focused on its risks to state laws without noting these elements.

Sen. Bernie Sanders (I-Vt.), in a Nation op-ed on April 7, 2026, warned of AI risks including job losses, privacy erosion, social isolation, environmental strain from data centers, increased war likelihood via robot soldiers, and existential threats from superintelligent AI. Sanders questioned whether billionaires like Elon Musk and Jeff Bezos, investing hundreds of billions (global AI private investment reached $252 billion in 2024 per Stanford HAI 2025 AI Index), seek power over democracy. These claims lack specific sourcing for trends like youth AI companions or trillion-dollar investments, and economists note AI may reshape rather than net destroy jobs, as in past tech shifts (NYT, HCAMag).

The Nation editorial claimed Sanders and Rep. Alexandria Ocasio-Cortez (D-N.Y.) proposed a late-March 2026 moratorium on data centers, but no confirming bill or press release appears on their official sites or Congress.gov. A February Economist/YouGov poll found 63% of Americans expect job losses from AI, while a Marquette Law School Poll (exact figures disputed) showed concerns over data center costs in Wisconsin. Environmental groups, including Greenpeace, warned in a December 2025 letter of disruptions from AI data centers' energy and water use.

Trump signaled deregulation at a summer 2025 Pittsburgh 'energy and innovation summit,' stating 'regulation be damned,' per the Nation editorial. First Lady Melania Trump reportedly described AI as 'personified' with a robot educator 'Plato' at a March 25, 2026, White House event, and Anthropic CEO Dario Amodei allegedly called AI a 'general labor substitute,' but these quotes remain unverified in public records. Sanders responded, per the editorial, opposing robot teachers and advocating decent wages for educators.

Proponents, including DoD and White House releases, stress AI's role in national security and innovation, with PSPs like AI swarms targeting China competition—details omitted in Nation pieces. State-level activity includes over 1,000 AI bills, creating a 'patchwork' that federal preemption addresses, per Morrison & Foerster analysis. Data center investments generated jobs and revenue, e.g., Virginia's Loudoun County taxes nearly covering its budget (City Journal, March 2026).

No active U.S.-Iran military conflict exists as of April 2026, contradicting The Nation's opening frame of AI as a tool in such a war; diplomatic references, like VP Vance on negotiations, prevail. Hegseth's title is officially 'Secretary of Defense,' though 'Secretary of War' appears on war.gov; standard media uses the former. The AI push formalizes partnerships with military-tech firms, but congressional oversight and laws-of-war compliance lack detailed guidance in the January memo, per critics.

AI leaders like Musk have called for regulation (Davos statements), countering unchecked-power narratives. Congress has not passed comprehensive AI legislation, with Sanders noting inaction. The Pentagon's strategy aims for AI in business practices and data processing within months, potentially centralizing decisions but enhancing speed.

Environmental impacts include data centers' electricity demands straining grids, as Sanders noted, though U.S. investments hit $750 billion in 2025 for skilled jobs. Privacy concerns arise from tracking capabilities, and robot soldiers could lower war thresholds, per Sanders' unverified question. The Quincy Institute authors (Abou-Elias, Hartung) advocate restraint from anti-interventionist views.

As of April 2026, the 'AI-first' directive and strategy proceed without major halts, amid calls for frameworks. DoD has not responded to Nation critiques directly. Polls indicate public wariness, with 69-73% expressing economic or speed concerns in cited surveys (figures flagged as approximate).

The integration reflects post-World War II tech-security equations, but past failures like Iraq/Afghanistan networked warfare highlight limits, as Abou-Elias and Hartung argued—though Gulf War outcomes included decisive victories despite munitions volumes. Full implementation timelines remain fluid, with PSPs targeted for months-long rollout.