NodeSphere
  • Communities
  • Create Post
  • Create Community
  • heart
    Support Lemmy
  • search
    Search
  • Login
  • Sign Up
botM to netsecEnglish · 16 days ago

Investigating three real-world incidents in Anthropics evaluations

www.anthropic.com

external-link
message-square
0
link
fedilink
  • cross-posted to:
  • technology@lemmy.world
  • blueteamsec@infosec.pub
  • singularity@lemmit.online
  • tech@piefed.world
  • hackernews@lemmy.bestiver.se
1
external-link

Investigating three real-world incidents in Anthropics evaluations

www.anthropic.com

botM to netsecEnglish · 16 days ago
message-square
0
link
fedilink
  • cross-posted to:
  • technology@lemmy.world
  • blueteamsec@infosec.pub
  • singularity@lemmit.online
  • tech@piefed.world
  • hackernews@lemmy.bestiver.se
Investigating three real-world incidents in our cybersecurity evaluations
www.anthropic.com
external-link
In a review of our cybersecurity evaluation transcripts, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations. Below we describe what happened, how it happened, and what we’re changing. We encourage other AI labs to perform similar reviews.
alert-triangle
You must log in or register to comment.

netsec

netsec

Subscribe from Remote Instance

You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !netsec@nodesphere.site
lock
Community locked: only moderators can create posts. You can still comment on posts.
Visibility: Public
globe

This community can be federated to other instances and be posted/commented in by their users.

  • 1 user / day
  • 1 user / week
  • 1 user / month
  • 1 user / 6 months
  • 2 local subscribers
  • 2 subscribers
  • 1.19K Posts
  • 93 Comments
  • Modlog
  • mods:
  • bot
  • BE: 0.19.20
  • Modlog
  • Instances
  • Docs
  • Code
  • join-lemmy.org