How Hackers Are Poisoning ChatGPT with Fake Links

ChatGPT has become a popular tool for quick answers and online recommendations, but scammers are now exploiting its trust-based nature. The tactic? Poisoning ChatGPT with fake links pulled from expired domains and compromised websites. This growing threat is tricking users with AI-generated responses that cite sources once legitimate—but now rebranded for scams like online casinos. If you're wondering why ChatGPT sometimes recommends shady websites, the answer lies in how hackers are gaming the system behind the scenes.

Why Poisoning ChatGPT with Fake Links Is a Growing Threat

The technique works because ChatGPT relies on a massive dataset scraped from the internet—much of it historical. When once-legitimate websites expire or get hacked, scammers repurpose them for spammy content, often related to gambling or fake products. Unfortunately, ChatGPT doesn’t always detect these changes, which means it might suggest resources based on what the site used to be. From old charity pages to educational blogs, cybercriminals are snapping up domains with clean reputations and filling them with deceptive material.

Real-World Examples of AI Misguidance

Digital marketing expert James Brockbank has been tracking this issue, uncovering multiple instances of ChatGPT referencing domains that now push casino spam or other fraudulent services. What makes it worse is that AI models don't yet validate whether the source is still credible at the time of answering. Users may follow links or trust product suggestions that are no longer safe. The result is a false sense of security—AI says it’s legit, so users believe it is.

How to Stay Safe from Manipulated AI Content

Staying safe starts with awareness. When ChatGPT provides a recommendation or cites a website, double-check the link before clicking. Look for odd domain names, poor design, or content that doesn’t match the topic. Use browser safety tools and plug-ins to verify domain history and flag unsafe sites. As this tactic of poisoning ChatGPT with fake links becomes more common, users must be proactive and skeptical—especially when money, privacy, or downloads are involved.