Davide Bragetti

Serving Europe.

HOME
CATEGORIES
TAGS
ARCHIVES
ABOUT
VISUAL STORIES

Home Tags debate

Tag

debate 1

A meditating monk in saffron robes at the centre of a temple veranda overlooking misty mountains; behind him a stack of virtue-labelled books (Ethics, Compassion, Non-Harm, Mindfulness, Wisdom, Right Intent, Interdependence, Patience, Equanimity, Right Action) is being consulted by a small grey creature reading from a Rogue Manual; three other grey creatures carry a log and tend a fire on the right, going about ordinary work. Allegory for safe AI coexisting with unsafe AI in a shared ecosystem.

Does Safe AI mean nothing bad can ever happen?

Even granting that mechanistic interpretability gets us to safe AI, does that guarantee a safe world? Notes from the BlueDot Unit 4 debate.

May 7, 2026 ai-safety

Recently Updated

Uno, nessuno, centomila e tutti
Misalignment by Reaction
Does Safe AI mean nothing bad can ever happen?
To Be or to Game
Unpacking Malware

Topics

ai-safety cryptography alignment bluedot governance malware-analysis malware mesa-optimization active-directory agent-autonomy

© 2026 Davide Bragetti. All rights reserved.

Topics

ai-safety cryptography alignment bluedot governance malware-analysis malware mesa-optimization active-directory agent-autonomy