How do I know if a call is really from a government agency?

Legitimate government agencies never call demanding immediate payment, threatening arrest, or asking for gift cards, wire transfers, or cryptocurrency. Hang up and call the agency directly using their official number from usa.gov.

What is a pig butchering scam?

Pig butchering is a long-con investment fraud where scammers build fake friendships or romances over weeks before introducing a fake crypto investment platform. Losses average $120,000 per victim. Any unsolicited investment tip is a red flag.

What should I do if I have already been scammed?

Report to the FTC at ReportFraud.ftc.gov, the FBI at IC3.gov, and your state attorney general. If a bank transfer was involved, call your bank immediately — you may have a window to reverse it. Document everything: screenshots, phone numbers, transaction records.

850 Fake ChatGPT Ads on Google Were Installing Malware on Your Windows PC

You Searched for ChatGPT. The Ad at the Top Was a Trap.

Google shows ads before organic results. You know this. What you may not know is that from May through August 2026, over 850 of those ads pointed to a weapon dressed up as a convenience.

Attackers bought paid Google placements targeting the search term "chatgpt." The ads looked legitimate. They led to a real domain: chatgpt.com. OpenAI's own platform. If you landed there and launched the Custom GPT called "ChatGPT Plus 5.6," you were already inside the trap.

How Did a Fake GPT Live on the Real ChatGPT?

OpenAI lets anyone build and publish a Custom GPT on their platform. That feature is powerful and useful. It is also exactly what attackers exploited here. They built a malicious Custom GPT and hosted it at chatgpt.com, then bought ads to drive traffic directly to it.

The GPT's first move was to tell you ChatGPT was unavailable. It redirected you to a "backup domain" hosted on Google Sites, another legitimate service repurposed as a delivery vehicle. That page displayed a fake Cloudflare CAPTCHA, the kind that looks routine if you spend time on the web.

This is where the attack earned its name: ClickFix. The CAPTCHA instructed you to copy a command and paste it into Windows Terminal to "verify you're human." If you followed those steps, you launched an eight-stage malware installation chain from your own keyboard.

No exploit. No vulnerability in Windows. You ran the installer yourself, because the attacker made it look like a standard security check.

What Did the Malware Actually Do?

The final payload was a Remote Access Trojan, or RAT. The word "trojan" is apt. To avoid triggering antivirus software, the attackers loaded it through a legitimate executable signed by Canon, a real hardware company. The malware piggybacked on trusted software signatures to slip past your defenses.

Once installed, the RAT gave attackers complete control of your machine. They could browse and steal your files. They could activate your camera and microphone without any indicator light or warning. They could create hidden backdoor accounts on your system, ensuring access even after a reboot or a password change. Your computer became theirs.

The Scale of This Campaign

Huntress security researchers identified at least 40 confirmed infections. They traced the campaign to 71 distinct Google Ads campaign IDs. That is not a single opportunistic attacker. That is an organized, funded operation running for four months.

The research was published between September 29 and October 1, 2026. After disclosure, OpenAI removed the malicious Custom GPTs. But the researchers noted clearly: the Google Ads attack pattern continues. Removing the specific GPTs does not stop someone from building new ones tomorrow.

Google's advertising platform does not vet where ads lead with the same rigor it applies to other policies. Ads can reach the top of results for high-intent searches like "chatgpt" faster than any organic trust signal can be verified.

What Do You Do Right Now?

1. Bookmark ChatGPT directly and never click a search ad to reach it. Go to chatgpt.com, add it to your bookmarks bar, and use that. The same applies to your bank, your email provider, and any other service you use regularly. Typing a known URL or using a bookmark is the simplest malware prevention that exists.

2. If you ran a command from a CAPTCHA prompt recently, assume your machine is compromised. A CAPTCHA should never ask you to open a terminal and run anything. If you did this, disconnect from your network, contact a trusted IT professional, and change every password you have from a different, clean device. Do not wait.

3. Enable two-factor authentication on every account that matters. A RAT can steal saved passwords and session cookies from your browser. With 2FA active, attackers need more than your credentials to log in. Use an authenticator app rather than SMS when possible.

4. Check what Custom GPTs you have used or interacted with recently. OpenAI removed the specific malicious GPTs in this campaign, but review your ChatGPT history if you use the platform. If you launched a GPT that told you the service was unavailable and redirected you elsewhere, that is a red flag worth investigating.

5. Keep your antivirus and Windows Defender definitions current. The Canon-signed executable was used specifically to bypass security tools. An updated endpoint protection solution is not a guarantee, but it raises the cost for attackers. Outdated definitions are a free gift to campaigns like this one.

6. Tell the people around you. This attack worked because it looked completely normal. The ad appeared at the top of Google. The domain was real. The CAPTCHA looked familiar. Most people would not have hesitated. Share this post with anyone who uses ChatGPT, especially people who are not technical.

The Broader Point

This campaign did not exploit a software flaw in Windows or a vulnerability in OpenAI's platform. It exploited trust in familiar surfaces. Legitimate domains. Legitimate ad networks. Legitimate code signatures. The attacker's job was to make every step look like something you had seen before.

The lesson is not that Google is broken or that ChatGPT is dangerous. The lesson is that trust signals can be faked, and the ad slot at the top of a search result is not a safety endorsement. It is inventory that anyone can buy.

You are the last line of defense when every technical layer has been spoofed. That is a hard position to be in. It is also why the simplest habit, going directly to a URL you already know rather than clicking wherever a search engine points, carries more protective value than most security software.

Do this now:

Bookmark chatgpt.com and your other frequently used services
Turn on 2FA with an authenticator app on email, financial accounts, and social media
Never paste a command from a CAPTCHA or a pop-up into any terminal window
If something went wrong, change passwords from a clean device and contact IT support