Two new analyses found that ChatGPT’s cited sources changed as its search traffic shifted between hidden retrieval pipelines. The findings. Research from Chris Green and Suganthan Mohanadasan adds a new complication to AI visibility tracking: the final answer doesn’t reveal how ChatGPT selected its sources. Both researchers found internal source-selection labels, including Labrador, Bright, Oxylabs, and SERP. Those labels sit behind the answer, not in the citation cards users see. Hidden pipelines changed sources . Green tested 1,000 prompts up to 10 times each and captured 9,946 completed search runs. Most prompts stayed on one retrieval…
This is a curated summary. The full story was reported by Search Engine Land.
Read the full story at Search Engine Land
