Google eyes Pentagon deal
Google is in talks with the Pentagon to deploy its Gemini AI in classified operations. The reporting says the draft contract language would bar use for domestic mass surveillance and would forbid autonomous‑weapons targeting without appropriate human oversight. (Google eyes AI deal with Pentagon to deploy Gemini AI in classified ops - BusinessToday, Google in talks with Pentagon over Gemini AI deployment - Information By Investing.com)
Google is negotiating with the Pentagon to let the Defense Department use Gemini inside classified systems, according to reports published April 17. (yahoo.com) The talks center on draft contract language that would still let the Pentagon use Gemini for “all lawful purposes,” while barring domestic mass surveillance and blocking autonomous-weapons targeting without human oversight. (businesstoday.in, economictimes.indiatimes.com) Classified deployment means Google’s models would run in secret government networks rather than only in public cloud tools. The report said the language is still being negotiated and no final contract has been announced. (yahoo.com, tradingpedia.com) The timing follows a broader Pentagon push to pull commercial AI into military work. In a January 9, 2026 strategy memo, the department said it wants to become an “AI-first” force and to unleash experimentation with leading U.S. models across the department. (media.defense.gov) Google’s position has also shifted. On February 4, 2025, the company removed an earlier public pledge not to build AI for weapons or surveillance, replacing it with principles framed around benefits, risks, international law, and human rights. (cnbc.com, blog.google) That change marked a break from the line Google drew after Project Maven, the Pentagon drone-imagery program that triggered an internal revolt in 2018. More than 3,100 employees signed a letter that year asking Sundar Pichai to cancel the work. (cnbc.com) Google now says its AI work is governed through testing, safeguards, and review processes embedded across the product lifecycle. In its February 18, 2026 Responsible AI Progress Report, the company said its AI Principles remain the “north star standards” for research, products, and business decisions. (blog.google) The draft Pentagon language appears designed to square those public commitments with military demand for classified access. If the deal is completed, Google would move deeper into the same national-security AI market that has pulled in other frontier-model companies over the past year. (quiverquant.com, businesstoday.in)