Discussion
Loading...

Post

Log in
  • About
  • Code of conduct
  • Privacy
  • Users
  • Instances
  • About Bonfire
Eva Wolfangel
Eva Wolfangel
@evawolfangel@chaos.social  ·  activity timestamp 2 months ago

Ich sags ungern, aber: Told you so.
Vor genau diesem Szenario warne ich seit Monaten in diversen Vorträgen, Podcasts, Texten. Oft kommt zurück: "Ach komm, jetzt mach doch nicht immer alles schlecht. Wir können nicht auf KI verzichten, es machen doch alle..." Und: "Diese Attacken gibt es doch nur in der Theorie."
Voila: Hier eine reale Attacke des chinesischen Geheimdienstes mit der Hilfe von KI Agenten von Anthropic.
https://www.anthropic.com/news/disrupting-AI-espionage

#aiagents #anthropic #chatgpt #cybersecurity

Screenshot aus dem Bericht:

They also told
Claude that it was an employee of a
legitimate cybersecurity firm, and
was being used in defensive testing.
The attackers then initiated the
second phase of the attack, which
involved Claude Code inspecting the
target organization’s systems and
infrastructure and spotting the
highest value databases. Claude was
able to perform this reconnaissance
in a fraction of the time it would've
taken a team of human hackers. It
then reported back to the human
operators with a summary of its
findings.

In the next phases of the attack,
Claude identified and tested security
vulnerabilities in the target
organizations’ systems by
researching and writing its own
exploit codes Having done so, the
framework was able to use Claude to
Screenshot aus dem Bericht: They also told Claude that it was an employee of a legitimate cybersecurity firm, and was being used in defensive testing. The attackers then initiated the second phase of the attack, which involved Claude Code inspecting the target organization’s systems and infrastructure and spotting the highest value databases. Claude was able to perform this reconnaissance in a fraction of the time it would've taken a team of human hackers. It then reported back to the human operators with a summary of its findings. In the next phases of the attack, Claude identified and tested security vulnerabilities in the target organizations’ systems by researching and writing its own exploit codes Having done so, the framework was able to use Claude to
Screenshot aus dem Bericht: They also told Claude that it was an employee of a legitimate cybersecurity firm, and was being used in defensive testing. The attackers then initiated the second phase of the attack, which involved Claude Code inspecting the target organization’s systems and infrastructure and spotting the highest value databases. Claude was able to perform this reconnaissance in a fraction of the time it would've taken a team of human hackers. It then reported back to the human operators with a summary of its findings. In the next phases of the attack, Claude identified and tested security vulnerabilities in the target organizations’ systems by researching and writing its own exploit codes Having done so, the framework was able to use Claude to
  • Copy link
  • Flag this post
  • Block
Eva Wolfangel
Eva Wolfangel
@evawolfangel@chaos.social replied  ·  activity timestamp 2 months ago

Nochmal hierzu: Ich suche schon länger für eine Recherche nach "echten" Cyberangriffen mittels generativer KI / KI-Agenten. Es gibt einige spannende Angriffe von Sicherheitsforscher:innen, aber ich habe bis auf den aktuellen noch keine "echten" Fälle gesehen. Also eben nicht, wie Cyberkriminelle mit Chatbots bessere Phishing-Mails etc schreiben, sondern automatisierte Angriffe wie hier beschrieben. Oder wie KI-Agenten eines Unternehmens gegen dieses selbst gerichtet werden. Freu mich über Tipps.

  • Copy link
  • Flag this comment
  • Block

bonfire.cafe

A space for Bonfire maintainers and contributors to communicate

bonfire.cafe: About · Code of conduct · Privacy · Users · Instances
Bonfire social · 1.0.1 no JS en
Automatic federation enabled
Log in
  • Explore
  • About
  • Members
  • Code of Conduct