Universal Music is launching an AI music platform with ElevenLabs14m ago OpenAI Introduces Data Agent in ChatGPT Work to Analyze Company Data16m ago Robotics Startup Skild AI Hits $100 Million in Revenue Run Rate22m ago Is AI The Silver Bullet For Project Portfolio Management?22m ago An Anthropic researcher just quit, saying OpenAI and Anthropic are 'gambling with our lives'24m ago Cohere Debuts Open-Weight 218B Mixture-of-Experts Machine Translation Model25m ago Bending Spoons agrees to buy Miro for $1.355bn, weeks after Airtable29m ago Winmau And Autodarts Bring Smart Scoring To Your Dumb Dartboard30m ago Typewise orchestrates customer-service AI agents36m ago Zuckerberg tells Sources labs holding back AI models is ‘quite dangerous’44m ago Nintendo's latest sale is 'made possible' by the tariff refunds it's hoarding51m ago Now everyone can put data to work52m ago Instagram now lets you add tagged photos to your profile grid52m ago How To Secure The AI Features Hiding Inside Your Sanctioned SaaS Apps52m ago

Universal Music is launching an AI music platform with ElevenLabs14m ago OpenAI Introduces Data Agent in ChatGPT Work to Analyze Company Data16m ago Robotics Startup Skild AI Hits $100 Million in Revenue Run Rate22m ago Is AI The Silver Bullet For Project Portfolio Management?22m ago An Anthropic researcher just quit, saying OpenAI and Anthropic are 'gambling with our lives'24m ago Cohere Debuts Open-Weight 218B Mixture-of-Experts Machine Translation Model25m ago Bending Spoons agrees to buy Miro for $1.355bn, weeks after Airtable29m ago Winmau And Autodarts Bring Smart Scoring To Your Dumb Dartboard30m ago Typewise orchestrates customer-service AI agents36m ago Zuckerberg tells Sources labs holding back AI models is ‘quite dangerous’44m ago Nintendo's latest sale is 'made possible' by the tariff refunds it's hoarding51m ago Now everyone can put data to work52m ago Instagram now lets you add tagged photos to your profile grid52m ago How To Secure The AI Features Hiding Inside Your Sanctioned SaaS Apps52m ago

Sam Altman admits OpenAI can’t control Pentagon’s use of AI

The GuardianPublished Mar 4, 2026

OpenAI CEO Sam Altman acknowledged that his company cannot regulate how the Pentagon applies AI technologies, highlighting challenges in controlling military use of AI. This admission underscores the limits of OpenAI's influence over external AI deployments.

More on this story

Facing Backlash, OpenAI Amends Pentagon Deal to Add More Anti-Surveillance Verbiage
Mar 3· 2 sources
Sam Altman in Damage Control Mode as ChatGPT Users Are Mass Cancelling Subscriptions Because OpenAI Is “Training a War Machine”
Mar 2· Futurism
OpenAI’s Sam Altman announces Pentagon deal with ‘technical safeguards’
Feb 28· TechCrunch
Our agreement with the Department of War
Feb 28· Open AI News

#ai-control #military-ai #openai #pentagon #sam-altman

More in AI Research

AI Research4 min read

An Anthropic researcher just quit, saying OpenAI and Anthropic are 'gambling with our lives'

Covered by 12 sources1d ago

AI Research3 min read

Worried Anthropic researchers warn that AI ‘could kill all humans’

Covered by 8 sources1d ago

AI Research4 min read

Suno trained its v6 AI music models with help from Warner and BMG

Covered by 5 sources1d ago