By Vincee Cole
Publication Date: 2026-08-19 14:32:00
You probably don’t think of your ChatGPT or Claude account as a security target. But hackers do. AI platform credentials are an increasingly active commodity on dark web marketplaces, and Axios confirmed that AI account hijacking — sometimes called “LLM hijacking” — has become a growing threat category in 2026. Here’s how to check if someone else is using your account, and what each platform does (or doesn’t do) to protect you.
Why Hackers Want Your AI Account
Your AI account contains more sensitive data than you might realize. Chat histories often include confidential business information, personal details, code snippets, financial data, and private conversations. A hacker with access to your ChatGPT account can read every conversation you’ve ever had with the model.
Beyond data theft, compromised accounts get resold for their subscription value. A stolen ChatGPT Plus or Claude Pro account grants the buyer premium AI access without paying. At scale, reselling stolen AI subscriptions has become a meaningful revenue stream for cybercriminals.
We’ve also seen AI accounts used as attack vectors. A researcher at security firm Zenity demonstrated how a compromised Claude in Chrome account could be used for unauthorized Amazon purchases, and a ChatGPT Atlas vulnerability allowed zero-click hijacking through planted comments on X. These attacks don’t require sophisticated hacking — they exploit architectural design flaws in how AI agents browse the web.
