As AI agents spread, so do the risks, scholars say

[HVNews]

By injecting a single altered image into a group of communicating chatbots, researchers theorize that up to a million such agents could be compromised in a short period of time.

Click here to view the article.