ChatGPT Just Became a Top-10 Phishing Target: Here’s What That Means for AI Users

For the first time in its history, OpenAI’s ChatGPT has joined the list of the world’s most impersonated brands in phishing attacks. According to Check Point Research’s Q2 2026 Brand Phishing Report — published July 23–24, 2026 — ChatGPT accounted for 1.1% of all brand phishing attempts globally during the quarter, making it the only AI service in the top 10. The report signals a clear shift: cybercriminals are now treating AI platforms as prime phishing real estate, and they are deploying targeted, subscription-anxiety scams at scale .

The Top 10 Most Impersonated Brands in Q2 2026

Microsoft continues to dominate brand impersonation by a wide margin, but the biggest story this quarter is the entry of an AI-native brand. The top five impersonated brands together account for more than half of all brand phishing attempts tracked .

Rank	Brand	Share of Brand Phishing Attempts
1	Microsoft	23.0%
2	LinkedIn	11.6%
3	Google	6.7%
4	Apple	5.8%
5	Amazon	5.2%
6–10	WhatsApp, Facebook, Instagram, Netflix, ChatGPT (ranked ~10th)	~1–3% each

Microsoft has held the top spot for multiple consecutive quarters, but LinkedIn’s rise to second place is notable and reflects an increase in professional-credential theft campaigns . ChatGPT’s entry is the headline change: it is the first time an AI platform has cracked the top 10 since Check Point began tracking brand phishing .

Why ChatGPT’s Top-10 Debut Matters

1. Attackers follow the user base — and AI has a massive one

The pattern is well-established: as a platform reaches hundreds of millions of users, it becomes a high-value phishing target. The same thing happened with Microsoft, Google, Facebook, and now ChatGPT . “ChatGPT has hundreds of millions of users, and attackers follow the user base,” as multiple cybersecurity analysts have noted .

2. Subscription anxiety is the new engine of phishing

The most common ChatGPT-themed phishing campaigns do not try to trick users with technical jargon. They exploit a simple, powerful fear: the fear of losing access. Attackers send urgent emails claiming that a ChatGPT Plus payment has failed or that a subscription is about to expire, and they direct victims to lookalike payment portals .

Real-world examples of ChatGPT-targeted phishing include:

• "Payment Failed / Update Payment Details" emails (July 2026) – MailGuard intercepted a wave of emails claiming the recipient’s last ChatGPT Plus payment failed, directing them to a fake Stripe checkout page designed to steal email addresses, credit card numbers, and CVV codes .
• "Payment Method Update" scam (February 2026) – A phishing campaign sent minimalist payment-notification emails threatening account suspension within 2 days unless payment info was updated .
• 4,500-email campaign targeting South Africa (May 2026) – Microsoft detected a ChatGPT-themed attack where 97% of recipients were in South Africa, using malicious URLs to harvest credit card and personal data .
• Fake subscription renewal emails with malicious SVG attachments (March 2026) – Attackers sent renewal reminders that linked to phishing portals or included SVG file attachments that executed scripts to redirect victims .

The psychological hook is always the same: “You will lose access to GPT-5, your workflows, or your paid plan” .

3. AI services are both weapon and target

There are two converging trends at work. First, attackers use generative AI to write more convincing phishing emails with proper branding, logos, and flawless grammar — making it harder for users to spot fakes . Second, attackers are now impersonating AI services themselves to steal credentials and payment data .

More concerning is a vulnerability called ChatGPhish, disclosed in June 2026 by Permiso Security. It demonstrated that any public web page summarized by ChatGPT can inject phishing links, fake security alerts, and QR codes directly into the ChatGPT user interface via indirect prompt injection — without compromising OpenAI’s infrastructure . A user who asks ChatGPT to summarize an attacker-controlled page could see a fraudulent “security alert” or “update your billing” prompt rendered as part of the trusted ChatGPT UI.

4. The trend is accelerating sharply

Kaspersky reported that ChatGPT-mimicking cyberthreats surged 115% in the first four months of 2025 compared to the prior period . The Q2 2026 ranking formalizes what security firms had been warning: AI services are now firmly on the radar of cybercriminals, and the volume of attacks is growing faster than most users expect.

How to Protect Yourself from ChatGPT-Targeted Phishing

• Never click payment links in emails. Always go directly to chat.openai.com/account/billing to check your subscription status .
• Check the sender address carefully. In one campaign, the sender was "info@openai-support.com" — a domain that is not owned or operated by OpenAI .
• Enable multi-factor authentication on your OpenAI account and any other major platform accounts .
• Treat urgent emails with skepticism. Any email that creates artificial urgency about losing access to an AI service is a red flag. The urgency is the scam’s engine .
• Be aware of the ChatGPhish vector. If you ask ChatGPT to summarize a web page you do not fully trust, be cautious of any links or alerts that appear in the response. Verify them independently from the source web page .