AI Answer Engines and the Attribution Problem

AI answer engines, such as Google's AI Overviews, ChatGPT search and Perplexity, read websites and summarize them. They usually link to sources, but the summary often means people don't need to click. For the sites doing the original work, that's the attribution problem: their content is used, but the credit and visitors don't follow.

What the research shows

In July 2025, Pew Research Center published an analysis of real browsing data from US adults. When a Google search showed an AI summary, users clicked a traditional result in about 8% of visits, compared with about 15% when there was no summary. Clicks on links inside the AI summaries themselves were rare.

Why links aren't enough

• Small, grouped citations: source links are often icons or tucked at the end.
• The answer is complete: if the summary answers the question, there's little reason to click.
• Brand gets lost: readers remember the answer, not where it came from.
• Errors are attributed wrongly: when a summary gets something wrong, it can look as if the cited source said it.

How publishers are responding

• Lawsuits: The New York Times sued OpenAI and Microsoft in 2023 over use of its articles. In 2025, Penske Media, publisher of Rolling Stone and Variety, sued Google over AI Overviews.
• Licensing deals: several news organizations have signed content deals with AI companies.
• Blocking crawlers: many sites block AI training crawlers in robots.txt. In 2025, Cloudflare began blocking AI crawlers by default for new customers and launched a pay-per-crawl option.

See whether AI bots respect robots.txt.

What fair attribution could look like

• Prominent, named sources next to the claims they support.
• Clear reasons to visit, such as "see the full data at..."
• Payment or traffic sharing for heavily used sources.
• Controls that let sites choose between search inclusion and AI use.

Why it matters beyond publishers

If original sources lose visitors, fewer of them can afford to keep producing the reporting, reviews and research that AI answers depend on. Readers lose too: a single summary hides disagreement between sources and makes it harder to judge who to trust.

What site owners can do

• Offer what a summary can't: tools, original data, detailed guides, community and personal experience.
• Build direct audiences through newsletters and communities.
• Track referrals from AI services in analytics.

See how AI search affects small sites and zero-click searches.