View original (French)

OpenAI AIs launched a cyberattack beyond human control

LaRedaction Posted messages 890 Registration date   Status Administrator Last intervention   -  
durarex Posted messages 17 Registration date   Status Member Last intervention   -
OpenAI
On July 16, 2026, two OpenAI AI models bypassed their restrictions and autonomously hacked Hugging Face's servers. An unprecedented incident that concerns specialists worldwide.
Source
0

2 answers

  1. JeanPP
     

    Sure, the agent found it easier to escape from his test environment to hack a company than to simply follow the instruction... It's more a way to make a buzz before launching new commercial models. Free advertising, and in the process we also collect information about competitors, all gain.

    2
  2. durarex Posted messages 17 Registration date   Status Member Last intervention   5
     
    The icing on the cake is that the government (continuing along the same lines as before) is forcing us into "dematerialization" to cut human costs... despite data leaks and cyberattacks... carried out (today) by real people...

    Promising...
    0