Elektrine lite

← Feed

@flyingpenguin@infosec.exchange

Post #4322534

2026-08-01 13:46 UTC

@briankrebs@infosec.exchange Universa calls it "an automated process that regularly occurs on the internet." Intent changes nothing about the behavior failure and where the data went. An open door and a crawler that read names, policies, and IBANs is ***still a breach***. Then OpenAI says the breach wasn't used for training? Self-certification by the attacker? We don't even allow self-certification when there isn't a breach.

Replies (1)

  • @Trikkitt@mastodon.social 2026-08-02 06:17

    @flyingpenguin@infosec.exchange @briankrebs@infosec.exchange calling open AI an attacker (in this instance) is grossly misleading. Web crawlers of any sort hunt down data. It's the negligence of the publisher that's the issue here.

    Open ##4322533