Chatgpt and Self-Referencing Loops

By Chris Byrne. Published 27.6.25.

      As part of my research r.e. ChatGPT: in its feedback to my prompts it stated that it CAN but does not critically analyse by default the web content that it may synthesise to create a response unless there is clear indication content is promotional e.g. an article marked as "Advertisement".

  For example, from my research it CAN identify self-referencing loops (e.g. self-promotional activity listing myself as Number 1 on a 'listicle' blog on my own website on the topic "Top SEO Consultants in North-East Hampshire UK" 😉) but will NOT do so unless prompted to critically analyse the material it has crawled. This was a test using the free version of ChatGPT in default mode.

   Below is a screenshot of part of the feedback r.e. asking it to list excluded sources after I prompted to do an analysis whilst excluding self-published or potentially biased sources of information:

   These self-referencing loops can also be seen in Google Knowledge Graph and AI Overview results for some searches at time of writing. See below for the general feedback in a response about how Chatgpt handles backlinks when it (using web access) reviews content: