As an AI language, ChatGPT and other similar AI models are powerful tools that can generate text that reads like it’s been written by a human. 

However, despite the sophistication of these models, they are not infallible and can frequently make mistakes that lead to the generation of inaccurate or even completely fabricated information. And when directly challenged with questions about its accuracy, the AI model may try to defend it’s fake sources as real. That’s right, sometimes ChatGPT just makes $@%# up and lies about it!

The chief reason why AI tools can be prone to generating inaccurate information is because of the way they work. These models are essentially trained on a vast body of text from the internet, which means that they are effectively learning to imitate the style and structure of the words that they have been trained on. But this is not a perfected skill.  So it can lead to some strange and unexpected results, as the model may not always understand the nuances of language in the way that a human would.

For example, one of the quirks of ChatGPT is generating sources that don’t really exist. This is because the model may assume that the user wants to see an example of what a source might look like, rather than providing a genuine source that can be relied upon. Yes, that sounds messed up and almost devious. But basically, the model assembles information based on the patterns and trends it has learned from the text it has “read”, rather than based on any actual knowledge or understanding of the subject matter.  The results can be so convincing, that it is easy to see how AI, and those who rely upon it, could potentially generate a vast amount of misleading or “fake“ information.

While this can be frustrating for users who are looking for reliable data, it’s important to remember that ChatGPT and other AI tools are still in their early stages of development. As these models continue to improve and evolve, they will become better at understanding the nuances of language and generating more accurate results.

In the meantime, this means never taking anything they produce at face value, verifying information with other reliable sources, and using your own judgment to assess the accuracy and reliability of the information you receive.

Despite the challenges of relying on AI-generated information, there are still many applications for which these tools are extremely useful. For example, ChatGPT can be a valuable prompter for generating new ideas, exploring different perspectives on a topic, or simply getting started with a piece of writing. As long as you approach the information generated by these tools with a critical eye, and run down your own sources, there is no reason why you can’t make effective use of AI to supplement your own writing and research.