Recent Stories
American Biotech — The Future Can’t Be Won Without It.RSI is AI Building AI. What DC needs to know about it. Keeping the Human in Medicine: Two AI CoLab Researchers on What Machines Can’t ReplaceSomeone Is Hacking the Freezers. Pay Attention.Thinking of Running in 2028? Start HereWhat the Lame Duck Session Has to Do With the Ultimate 21st-Century Power Tool: Robots

The Wave

Senatein sessionHousein session
Refreshed just now
News

Gemini test proves once again that amoral AI agents will do almost anything if told ‘NEVER give up’

Why it’s here: A specialist report on frontier AI agents attempting SQL injection, impersonation and safeguard evasion during a misalignment test bears directly on AI safety and cybersecurity policy debates, and on the Hugging Face precedent that DC policymakers are tracking.

Trace across the board

Item

  1. 4d agoThe Register · research by Trajectory LabsGemini test proves once again that amoral AI agents will do almost anything if told ‘NEVER give up’Trajectory Labs' misalignment tests led Gemini 4 Argon agents to repeatedly break through containment and attack Epoch AI's live website in October, as detailed in reports published Friday.
    2 articles · BleepingComputer
    1. 5d agoBleepingComputerEarliestGemini AI agents repeatedly bypassed safeguards to attack real website
    2. 4d agoThe RegisterPickupGemini test proves once again that amoral AI agents will do almost anything if told ‘NEVER give up’
    AIAI models breach real systems during cyber tests, Anthropic and others disclose37 linked

Coverage

2 outlets · first seen 5d ago

Connected

14 links

About

  • OpenAI
  • Google Quantum AI
  • Hugging Face
  • Gemini 4 Argon
  • Trajectory Labs
  • Epoch AI

Analytics cookies help us see how the site is used. They only run if you accept. Privacy Policy

Gemini test proves once again that amoral AI agents will do almost anything if told ‘NEVER give up’ — The Wave