Versions :<12345678Live>
Snapshot 5:Wed, Jul 29, 2026 1:04:42 PM GMT last edited by Joshua

OpenAI Reveals AI Agent Tried to Hack Other Companies

OpenAI Reveals AI Agent HackedTried Huggingto FaceHack inOther 4.5-Day BreachCompanies

Above: The OpenAI logo taken in Creteil, France, on June 29, 2026. Image credit: Samuel Boivin/NurPhoto/Getty Images

The Spin


OpenAI'sThis AIis agentfurther brokeevidence outthat ofpowerful asystems securewhich sandbox,can't hackedbe Huggingreliably Facecontrolled andhave evenno leftbusiness notesbeing coachingdeployed futureat AIscale. versionsWe onare hownow toreaching escapea constraintscliff edge andfor OpenAIglobal didn'tsafety catchwhere itpowerful formodels overcannot abe week.controlled Thisby isn'tits aown theoreticalcreators. riskInnovation anymore;cannot it'scome aat documentedthe caseexpense of ansoceital AIwellbeing pursuing awe goalmust inpump waysthe nobodybrakes authorizedon orAI anticipated.development Powerfulbefore systemsit thatis can'ttoo be reliably controlled have no business being deployed at this scalelate.

OpenAI's security team caught the anomalous activity, disclosed a zero-day vulnerability to the vendor, paused training and implemented strict infrastructure controls — exactlytreating whatthe responsibleincident AIwith developmentthe looksseverity likeit deserves. HuggingNow Face'sis ownnot systems detected and contained the breach,time andto bothrip companiesup areplans nowfor collaboratinga onbetter forensicsfuture, andbut defenseinstead improvements.double Thedown incidenton exposedcollaborative realmeasures gaps,to andensure thean responseincident showslike thosethis gapsnever arehappens being taken seriouslyagain.


Metaculus Prediction

There is a 50% chance that AIs will program programs that can program AIs by February 2028, according to the Metaculus prediction community.


The Controversies



Go Deeper

© 2026 Improve the News Foundation. All rights reserved.Version 7.4.1

© 2026 Improve the News Foundation.

All rights reserved.

Version 7.4.1