An individual tested how three AI search engines—ChatGPT, Gemini, and Perplexity—cite sources for product recommendations. The experiment, conducted over two weeks in October 2026, involved asking fixed buyer questions and logging the results. A key finding was that the engines shared very few sources, and Gemini's source lists changed significantly between weeks, suggesting that repeated, standardized checks are necessary for accurate monitoring. The author developed a reproducible, albeit manual, method for this logging process. AI
IMPACT Highlights the variability in AI search engine source attribution, underscoring the need for consistent testing methodologies.
RANK_REASON User-generated content discussing methodology for evaluating AI search engine source citation.
Read on Mastodon — fosstodon.org →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →