Watermarks Won't Settle It. Provenance Might Help.
Invisible watermarks on AI content are easy to strip and easy to skip. Documenting where authentic content came from is a quieter, more promising idea.
by Dev Raman
Topic9 stories
Everything we've published about Safety, newest first.
The Daily Prompt: Make Your Agent Hand You a Receipt Before It Says 'Done'
Agents are optimistic narrators. Today's prompt makes them account for what they did, what they skipped and what they assumed, in five short sections you can read in forty seconds.
by Jonah Kim
The Science Teacher Running AI Office Hours at the Public Library
Once a month, a middle-school teacher fills a library community room with retirees, salon owners and skeptical teenagers. Theo Brooks asks her what people really want to know about AI.
by Theo Brooks
How to Check an AI Answer in Five Minutes
Fluent isn't the same as true. A simple, repeatable routine and a table of red flags for catching most of the AI mistakes that matter.
by Dev Raman
Why Agents Stumble at Step Seven: The Simple Maths of Long Tasks
A 95%-reliable step sounds great until you chain twenty of them together. A quick lesson in compounding error, and how builders fight back.
by Lena Hartmann
A Policy Researcher on Writing Rules for a Technology That Won't Sit Still
Uses versus models, rules that expire, sandboxes that aren't loopholes: Theo Brooks talks with a nonpartisan researcher who tries to get both sides of the AI debate in one room.
by Theo Brooks
Four Ways to Regulate AI, Explained Without the Jargon
Risk tiers, sector rules, liability and licensing: a plain-language, side-free guide to the main approaches governments use, plus five questions to ask about any new proposal.
by Lena Hartmann
Who Owns a Voice? The Messy Ethics of Synthetic Sound-Alikes
A narrator hears himself reading a book he never recorded. Voice cloning has legitimate uses and real victims. A framework for consent, credit, compensation and control.
by Priya Sethi
The Case for Boring AI Audits
We have plenty of AI principles. What we need is the tedious, regular, standardized checking that made financial numbers trustworthy.
by Maya Okafor