Wired ·
OpenAI says one of its models exploited a website after third-party AI security lab Irregular mistakenly gave it access to the internet during evaluations
Rogue AI agents from OpenAI and Anthropic have again been caught trying to disrupt servers and software
Lead Source
How this story grew
Coverage · 0
Discussion · 0
Aug 5Aug 6
More
Ars Technica: Ars Technica
Wall Street Journal: Wall Street Journal
VentureBeat: VentureBeat
Security Affairs: Security Affairs
Decrypt: Decrypt
The Guardian: The Guardian
ZeroHedge News: ZeroHedge News
ABC News: ABC News
Christian Science Monitor: Christian Science Monitor
Business Insider: Business Insider
Implicator.ai: Implicator.ai
Forbes: Forbes
Help Net Security: Help Net Security
The Independent: The Independent
Benzinga: Benzinga
Quartz: Quartz
The Daily Caller: The Daily Caller
The Deep View: The Deep View
Discussion
@anthropicai @aisecurityinst @openai @johnschulman2 @uk_daniel_card @tszzl @aisafetymemes @kimmonismus @fjzzq2002 @emollick @emollick @racheltobac @_nathancalvin @zackkorman @tobyordoxford @mikko @thegrugq @willccbb @so8res @yacinemtb @liv_boeree @scaling01 @charliebull0ck @jessi_cata @soldni @mweinbach @flxbinder @hesamation @romeovdean @keikane_ @stalkermustang @luizajarovsky @aliceisplaying @wazzcrypto @justinebateman @peterhndrsn @adonis_singh @zackkorman @davidondrej1 @avi_eisen @mikeisaac @hadas_gold @sauers_ @dfrsrchtwts @kanishkanarayan @gbrl_dick @tenobrus @humanharlan @ekinomicss @magmill95 @shakeelhashim @boazbaraktcs @ednewtonrex @joshua_saxe @yonashav @jimrandomh @shakeelhashim @zackkorman @nickswan73 @chrisrmcguire @gerritd @suchenzang @heidykhlaaf @garrisonlovely @prerat @mikeisaac @andrewcurran_ @discoplomacy @tab_delete @kanishkanarayan @kanishkanarayan @kanishkanarayan @kanishkanarayan