TL;DR Summary of How AI Assistants Differ in Citing Website Sources
AI assistants vary significantly in the websites they cite due to differences in training data biases and partnerships. Systems like ChatGPT, Perplexity, and Google AI Overviews each prioritize different types of sources, with only a few sites overlapping across all platforms. Citation methods differ, with some using real-time search and others relying on curated lists or licensed domains, impacting source diversity and accuracy.
Optimixed’s Overview: Understanding Source Citation Variations Among AI Assistants
Overview of Citation Differences in AI Assistants
AI assistants do not uniformly cite website sources; their approaches are shaped by their data training, partnerships, and technical methods. This results in varied source mentions even when addressing similar queries. Key factors influencing these differences include the presence or absence of real-time search, licensing agreements, and the handling of hallucinated content.
How Different AI Platforms Choose Their Sources
- Perplexity: Uses real-time web searches to find, rank, quote, and cite the best matching sources, resulting in a broader and more international set of websites, including many local and regional brands.
- ChatGPT (default): Lacks direct citation capability and may hallucinate sources, as it does not have real-time browsing by default.
- ChatGPT (with browsing): Conducts searches filtered for trusted or licensed domains, then cites those sources, often favoring media and publisher sites, especially in news and entertainment.
- Google AI Overviews: Relies on a search index combined with a curated citation list, emphasizing authoritative health, finance, encyclopedic, and Google-owned properties, but tends to exclude media, entertainment, and ecommerce sites.
- Bing Copilot: Uses Bing search to highlight snippets and includes citations based on those results.
Impact of Biases and Technical Limitations
Various biases affect which websites are cited, including:
- Partnership and licensing agreements that bias toward certain publishers.
- Technical factors like websites blocking AI bots or rendering JavaScript, limiting accessibility.
- Differences in regional and international focus, especially with Perplexity citing more localized brands.
Overall, top websites remain consistent as authoritative sources, but AI assistants differ in how they select and present citations, reflecting their underlying data, partnerships, and design choices.
Source: Data & Studies – SEO Blog by Ahrefs by Patrick Stox. Read original article.
