This website uses cookies

Read our Privacy policy and Terms of use for more information.

Sponsored by

Welcome Automaters, 👋

ChatGPT has a new problem, and it’s a big, sneaky, "where did my money go" kind of problem. 

A brand-new investigation by consumer safety research group Ask Silver has discovered that scammers are quietly slipping fake shopping websites directly into ChatGPT's search results. And the absolute worst part? The AI is handing them out to users like free candy at a parade.

Here’s exactly what happened: When Ask Silver researchers asked ChatGPT for information on super popular Russell & Bromley purses and bags, the AI cheerfully served up product details, prices, and a few links. But buried deep inside those sources were two completely fraudulent Russell & Bromley websites.

These clone sites looked incredibly polished, offered tempting discounts, and were about as real as a three-dollar bill. If you clicked them, you were walking straight into a digital trap. 

Ask Silver's Anna Jones flagged something particularly alarming about this mess: ChatGPT's underlying model may have been "poisoned."

In plain English, scammers flooded the internet with cloned, fake retail pages, which the AI then scraped and learned from during its training sessions. Now, when you ask ChatGPT for innocent shopping advice, it genuinely believes these fake sites are legit and points you straight to the fraudsters.

Louise Baxter, the literal head of scams at National Trading Standards, put it as bluntly as possible: do not assume a website is real just because an AI recommended it. Full stop! The bot does not actually know what is real.

According to researchers, this scam is especially effective when a brand unfortunately closes down or gets bought out. This leaves a massive gap with no official website, and fraudsters jump into that empty space fast.

How to spot these AI-approved scammers:

  • Their fake sites usually have messy URLs stuffed with desperate words like "official" or "deals."

  • They almost always demand payment by direct bank transfer, which is a shady method that’s nearly impossible for your bank to reverse once the money vanishes.

🛑 So What’s OpenAI Doing About It?

Well, OpenAI confirmed that the fraudulent Russell & Bromley sites flagged by the study have since been scrubbed from its search index. They also reminded everyone that users can report suspicious links directly through a dedicated reporting form.

But let's be real: taking down one or two fake sites does not fix the entire pipeline. Scammers are playing whack-a-mole with AI, and right now, the bots are losing.

P.S. We go even deeper on YouTube.

So go over there, hit Subscribe, hit the notification bell and come hang with us where the real conversation happens! 

Here’s yesterday’s breakdown: 

Here's what we have for you today

🔒OpenAI Launches ChatGPT Lockdown Mode to Fight Prompt Injection Attacks

OpenAI quietly dropped something massive on June 5th. It’s called Lockdown Mode, a brand-new optional security setting for ChatGPT, and you need to know about it. 

For one, it’s specifically built to defend against prompt injection attacks. If you aren't familiar, that’s basically when sneaky, chaotic bad actors hide secret instructions inside webpages, praying they can trick your AI into doing things you absolutely did not ask for.

And here’s the genuinely wild part: as ChatGPT has gotten way smarter at browsing the internet and pulling in outside information, hackers spotted a major loophole. They actually started planting invisible commands in online content, just waiting for an unsuspecting AI to stumble across them. 

So, what does Lockdown Mode actually do?

It completely slams the door on the trickiest attack routes. Once you switch it on, ChatGPT goes into total isolation. It cannot fetch images from the internet, download files, run Deep Research, or use Agent Mode. (Don't panic, though: you can still upload your own files and photos manually).

OpenAI is being totally real about the fact that this feature is not for everyone. It’s explicitly built for people and organizations sitting on highly sensitive data who need an extra layer of armor against information theft.

Crucially, it will not stop prompt injections from showing up in the content ChatGPT reads. Let's be clear about that. What it does do is block attackers from actually walking away with your data by severely restricting the network requests they would normally exploit. Huge difference.

Oh, and a quick bonus drop: OpenAI also rolled out an active session manager. This lets you see every single device currently logged into your ChatGPT account so you can boot them off, one by one or all at once.

How to get your own digital armor? Well to turn Lockdown Mode on: 

  • Head to your ChatGPT Settings

  • Tap Safety and Security

  • Then hit Advanced Security and toggle on Lockdown Mode

You can also temporarily disable the additional protection by selecting Manage from the status message that appears above the chat window and selecting Turn off for this chat

And don’t forget to meet us on YouTube so you can get our full take on this later today!

Your Website Shouldn't Look Like It Was AI-Generated

You know the pain. Your AI builder ships something functional. Technically a website. But your customers can tell.

Readdy.ai generates designs that actually convert, with clean layouts, professional copy, real visual hierarchy, and distinct custom designs that don't look like AI.

Get a site that's as special as your business. Live in under 5 minutes.

Agency quality. DIY price.

🧱 Around The AI Block

🤖 AI Workout Of The Day: How to Instantly Build a Complete Online Course with AI 

Want to create an online course fast without sacrificing quality? AI can help you build an entire course outline. From outlining modules to structuring content, suggesting resources, and breaking complex topics into digestible lessons, this prompt turns your subject knowledge into a full-blown course blueprint.

Whether you're teaching a skill, exploring a niche topic, or building a curriculum for your audience, this prompt turns your subject into a well-organized, engaging learning experience.

Some Tips for Maximum Output:

  • Be specific about your subject: Include the level (beginner, intermediate, advanced), the target audience (students, professionals, hobbyists), and the purpose (certification, self-paced learning, onboarding, etc.).

  • Add your flavor: Want a casual tone? Academic? Visual-heavy? Mention your preferences so the course fits your vibe.

  • Ask for optional formats: PDFs, videos, quizzes, slides, etc.

  • Use the output to build immediately: Upload into a course platform (like Teachable, Thinkific, or Notion), and start building right away.

💡 Prompts to try:

You are an expert in online course creation and instructional design. I want you to help me create a full course on:

 1. Subject: [Insert topic — e.g., “Freelance Writing for Beginners”]
 2. Level: [Beginner / Intermediate / Advanced]
 3. Audience: [Who is this for? E.g., college students, busy professionals, creators, etc.]
 4. Learning Goal: [What should learners be able to do by the end of the course?]

Please do the following:

 . Provide a compelling course overview.
 . Generate a table of contents with chapters/modules and clear titles.
 . Break each chapter into lessons or sections, focusing on key learning points.
 . Include a brief summary at the end of each chapter.
 . Recommend further reading or resources (websites, articles, videos) to reinforce the topic.
 . Optionally, suggest interactive elements like quizzes, discussion questions, or exercises.

Also make it engaging, structured, and easy to follow; something I can immediately turn into an online course.

Is this your AI Workout of the Week (WoW)? Cast your vote!

Login or Subscribe to participate

That's all we've got for you today.

Did you like today's content? We'd love to hear from you! Please share your thoughts on our content below👇

What'd you think of today's email?

Login or Subscribe to participate

Your feedback means a lot to us and helps improve the quality of our newsletter.

Reply

Avatar

or to participate

More From The Automated