Recently, SEO expert James Brockbank found in testing that the AI tool ChatGPT may reference unreliable information sources when providing business recommendations, including hacked websites and expired domains. This phenomenon has drawn attention from industry professionals, reminding users to verify the recommended information when using AI tools.
In the test, Brockbank found that ChatGPT sometimes references tampered or misleading content when recommending certain companies or services. These contents often appear on high-authority websites that have been hacked, or on expired domains that have been re-registered and used to publish irrelevant information. Since these domains originally had a high "authority," ChatGPT still considers them as credible sources when generating recommendations.
Brockbank pointed out that there are mainly two issues with this situation: first, hacked websites, which may have maliciously implanted information unrelated to the original content; second, the reuse of expired domains, where some organizations or individuals buy these domains and publish content unrelated to the original theme, misleading AI systems by leveraging their original search engine authority. He said, "Rebuilding a domain just to have a language model cite it? That's manipulation."
This discovery has far-reaching implications for businesses and users. For businesses, AI recommendations have become an important channel for brand exposure. However, if the recommended content comes from unreliable sources, it may mislead consumers and even damage brand reputation. For users, blindly trusting AI recommendations could lead to errors in information judgment.
Therefore, Brockbank called on AI developers to enhance their ability to identify and filter manipulated content, thereby improving the system's information screening capability. He also advised users to verify the sources of the information recommended by AI tools like ChatGPT, rather than fully accepting them. "We haven't reached the stage where we can completely trust ChatGPT's recommendations without verification," he emphasized.
With the rapid development of AI technology, ensuring the reliability of information has become an important issue. While users enjoy convenience, they also need to remain vigilant to avoid potential misinformation and risks.
Key Points:
🌐 ChatGPT may reference hacked websites and expired domains in business recommendations, making the information sources unreliable.
🔍 Users should verify the information recommended by AI rather than blindly trusting it to avoid misjudgment.
🚀 Experts call on AI developers to strengthen the identification and filtering mechanisms for manipulated content, enhancing information screening capabilities.