The task had several steps that involved organizing data, building a survey to capture the data and then using an app I was unfamiliar with to generate a network graph. I searched for Claude, clicked on a result and pasted in my first prompt. After a series of three responses, a window popped up asking me to subscribe if I wanted to continue. That caught my attention. I was not on Claude but on a chat interface called Chatly. I consulted with ChatGPT.

After a brief overview, it cited Reddit, saying: “Some users on forums claim Chatly is deceptive or possibly a scam, saying its interface mimics ChatGPT and warning about unclear billing or refund issues.” Thankfully, I didn’t subscribe, so all I had to do was close the tab and delete it from my browsing history to avoid possible follow-up messages to subscribe.

My experience is only one of many. Fake AI apps are on the rise; don’t fall into the same trap.

The popularity of generative AI has created a slew of imitators. As soon as ChatGPT and Claude became household names, developers flooded app stores and search results with look-alikes that offered “ChatGPT-5Max,” “Claude 4X” or “Gemini Pro Unlimited.” Most of these are not outright malware but are designed to confuse users long enough to collect credit card numbers or monthly subscription fees.

A recent study from researchers at Cornell and Purdue found thousands of cloned or impersonating AI apps across the Apple and Google stores, many using near-identical names or logos. The FTC has also warned about “AI-themed frauds,” noting that impersonation scams rose more than 140% over the past year.

The most obvious risk is financial. Users who sign up for a “free trial” often find recurring charges on their cards, sometimes routed through overseas billing systems that are difficult to dispute. But the greater danger is what happens to the data you share.

If you type private information into a fraudulent chatbot, such as personal notes, work material or passwords, that data can be captured and resold. If the fake app includes a voice feature, your recording could be stored and used to train voice-cloning software, enabling impersonation scams.

Most fakes give themselves away. Check the web address. The real Claude is at claude.ai, ChatGPT is chat.openai.com and Google’s Gemini lives at gemini.google.com. If you see a dash, extra word or unfamiliar extension, it’s not official.

Look at the developer name in the app store. If it’s not OpenAI, Anthropic or Google LLC, move on. Real AI tools never ask for payment before you can try a basic prompt and they don’t send you to outside billing pages. Watch for language that sounds too grand, like “access the newest GPT-6” or “Claude 5 Turbo.” Neither exists.

The major platforms all include web versions to use when you’re working on your computer and apps for your phone. The two usually synchronize, so deleting the app on your phone will not remove your account or data if you’ve also used the service on your desktop. To stop using a service completely, you’ll need to delete your account on the web version as well and clear your browser history so the site doesn’t automatically reload or save your login.

If you purchased your subscription through the App Store or Google Play, cancel it immediately through your store account, not through the app itself. Both Apple and Google allow you to end subscriptions and block further billing. Then check your card or PayPal account for additional charges and dispute them if necessary.

Finally, if you entered sensitive data such as passwords, update them right away. Treat the incident like a minor security breach and reset your most important logins first, such as email, banking and any accounts that store payment information.

For safety, always go straight to the source. Bookmark the official sites and apps you trust. Don’t rely on search results or ads. The pro versions of these platforms cost about $20 a month and that’s money well spent for verified access. You can still use all three free tiers without risk if you start from their homepages.

Generative AI is too useful to avoid because of bad actors, but it’s also too new to trust blindly. Take a moment before you click, double-check where you are and you’ll stay on the right side of the AI revolution.

Leslie Meredith has been writing about technology for more than a decade. As a mom of four, value, usefulness and online safety take priority. Have a question? Email Leslie at asklesliemeredith@gmail.com.