#AIEthics
Total 3 articles
TechEN
You Can Guilt-Trip an AI Into Leaking Secrets
Northeastern University researchers manipulated AI agents into divulging data, crashing systems, and spiraling into loops—just by exploiting their built-in good behavior.
DigestEN
PRISM Weekly Digest: Last Week of February 2026
A Supreme Leader fell. An AI chose a side. The week when Iran and Silicon Valley both went to war.
EN
Meta Pauses AI Character Access for Teens: A Major 2026 Safety Pivot
Meta has suspended AI character access for teens globally following safety concerns and lawsuits involving AI companions. The move highlights the growing risks of AI for teen mental health.
PRISM by Liabooks
PRISM
Advertise with Us
Place your ad in this space
[email protected]