Skip to content

Friday Deep-Dive · Friday, October 23, 2026

What a week, my AI horror story from the past seven days
The Fourth Estate AI Brief

Was this forwarded to you? Get The Fourth Estate AI Brief in your inbox →

Trevor Slette

Trevor Slette

Co-founder, Quadd.ai · Reply to me at trevors@quadd.ai

View this issue in your browser

Friday Deep-Dive

What a week, my AI horror story from the past seven days

I had to reset my computer twice in one week. I mean, 40-50 hours of my life wasted in keys, passwords and webhooks. That's what pushed me to lock down the AI tools I use every day. Here's the checklist I wish I'd had: three simple rules, the settings that matter in Claude, ChatGPT and Codex, and what to do if something goes wrong. Most of it takes no technical skill at all.

The AI lockdown checklist

AI assistants can now click, download and run things on your computer. That's useful, and it also means a bad website or a booby-trapped document can try to trick them. Here are the three rules I follow and the settings I use in each tool, written for people who are new to this.

One tricked AI with access to your email, website or Stripe account can do real damage in seconds. A few minutes of setup closes most of those doors.

Read the full post →

What is?

Prompt injection

What it is: Hidden instructions planted in a web page, email or document that try to trick an AI assistant into doing something you didn't ask for, like downloading a file or sending your information somewhere. The AI can mistake those instructions for yours.

Why publishers care: Newsrooms feed AI tools a constant stream of outside material: press releases, emailed tips, PDFs, websites. Any of it could carry hidden instructions, so a person approving each risky action is your best defense.

Visit ncsc.gov.uk →

None of this makes AI risk-free. Even the companies that build these tools say prompt injection isn't solved. But approving every risky action yourself, keeping passwords and keys away from the AI, and using a separate browser profile will stop most of the trouble. Menus change often, so check each step on your own screen. Trevor Slette runs Quadd.ai — AI tools built for publishers.

Trevor

Trevor Slette

Co-founder, Quadd.ai · 28-year community publisher

trevors@quadd.ai · Book a 15-min call · LinkedIn

P.S. Do one thing today: list every password and key your AI tools touch, because you can't replace what you don't know you have.

Found this useful? Forward to a colleague →

AP style without the AI mush.

Quadd's AP-Style Proofreading flags grammar, style, and headline issues with accept-or-reject suggestions. It tightens copy without inventing facts, flattening your voice, or turning every story into AI mush. The accept-or-reject part matters: you stay the editor. The AI just catches what a tired second pair of eyes would catch, if you had a second pair of eyes.

See how proofreading works →

Get the next one in your inbox

Subscribe — free, twice weekly