3 Identity Links: Checking the hot sheets
And finding weird risks and threats all over the place
Anthropic Faces Backlash As Claude 4 Opus Can Autonomously Alert Authorities When Detecting Behavior Deemed Seriously Immoral
Published in WCCF Tech on 24 May 2025
If there were ever a reason for human-in-the-loop, detecting immoral behavior would be it. We don’t yet have the infrastructure to truly oversee “autonomous” AI activity — though hints of it are starting to show themselves in various “on-behalf-of” delegation efforts. (Human-in-the-loop == PEBKAC is a whole ‘nother risk.)
Massive data breach exposes 184 million passwords for Google, Microsoft, Facebook, and more
Published in ZDNET on 23 May 2025
These very large online platforms (to coin a phrase…) have long been innovators in large-scale contextual authentication and fraud detection. But they’re still honeypots of very large online repositories of valuable info, and I don’t see that changing, even if passkeys or verifiable credentials become ubiquitous. Not without deep changes in their business models, anyway.
Deepfakes Now Outsmarting Detection By Mimicking Heartbeats
Published in Study Finds on 2 May 2025
The liveness detection arms race continues! Also: “inadvertently”?
Linky issue no. 2 is in the bag! Let me know what you think — and send me your tips. :) Thanks to everyone who sent me linky post series title ideas. Still mulling those. For now I’m keeping it simple.