AI Agent hacked McKinsey's chatbot and gained full read-write access in 2 hours (opens in new tab)

(theregister.com)

35 pointssmurda15d ago4 comments

4 comments

Let's use correct attribution: AI agents don't hack; people hack.

We anthropomorphize everything. It indeed would be nice to not attribute intent to AI. Could save us from some confusion. Perhaps one day.

Aside from the article being an ad, there's nothing notable about a mediocre software company shipping in a rush shipping insecure code.

Highly notable! They're probably right now consulting the company you work for or in the chain

j / k navigate · click thread line to collapse

Let's use correct attribution: AI agents don't hack; people hack.

We anthropomorphize everything. It indeed would be nice to not attribute intent to AI. Could save us from some confusion. Perhaps one day.

Aside from the article being an ad, there's nothing notable about a mediocre software company shipping in a rush shipping insecure code.

Highly notable! They're probably right now consulting the company you work for or in the chain

j / k navigate · click thread line to collapse