Yesterday, “Google Security” called me.

By sadlak · September 16, 2026 · Curated by George's Blog

Yesterday, “Google Security” called me.

Except it wasn’t Google.

'Ethan Baker' sounded professional. He knew my name, email and phone #.

And while we spoke, Google security emails appeared in my inbox.

For about a minute, it all looked legit.

Then I thought: I saw no problem before this Ethan called. And since when Google calls you?

So I said, “I’ll investigate it myself,” hung up and blocked the number.

I asked ChatGPT and Claude to reconstruct this phishing attempt:

1. The scammers created a random Google account.

2. They entered my email as its recovery address.

3. That triggered genuine automated emails from Google to verify.

4. They abused an “App password” alert to insert a fake case number, rep's name and support link.

5. Those emails were arriving during that short call, borrowing Google’s credibility.

Clever. And disturbingly simple. I can see why it works.

I reviewed my devices, connected apps and recovery settings, removed old access and strengthened the account. No evidence they got in.

The rule I’m keeping:

⚠️ If someone contacts you about an urgent security problem, end the conversation. Then open the company’s official app or website yourself.

• Never share a code.

• Never approve a login.

• Never follow the caller’s link.

What stayed with me was the 𝐛𝐨𝐫𝐫𝐨𝐰𝐞𝐝 𝐚𝐮𝐭𝐡𝐨𝐫𝐢𝐭𝐲:

because the message appeared inside Google, it felt like Google.

At ARTAN AI, we see the business version of this problem.

When Amazon, a brand’s website and retailers disagree (or marketplace data changes silently), ChatGPT and Alexa don’t ask the brand which detail is correct.

⚠️ They infer the answer.

At purchase intent, a wrong answer isn’t merely a data problem. It’s lost revenue.

We catch inconsistencies and silent changes early, align product details across sources, and make the correct answer easier for AI to understand and recommend.

Hopefully, this helps someone else hang up 30 seconds earlier. And if AI is misreading your products, that’s a call I’m happy to take! DM me.

View the original post on LinkedIn

More from sadlak