Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Yes, a 2026 study found sources classified as AI-generated among citations returned by Perplexity—but it did not show that those pages were necessarily inaccurate or spam. The study’s often-quoted figure of about 16% combines four AI search engines and applies only to cited sources researchers could successfully scrape. Separate audits raise questions about whether some citations support specific claims, but their findings are limited to their own samples and methods.
Is Perplexity citing AI-generated sources?
In a 2026 audit, researchers Mowafak Allaham and Nicholas Diakopoulos submitted 712 English-language, human-written queries about politics, health, and the environment to ChatGPT, Copilot, Gemini, and Perplexity. They collected the sources each system cited, scraped accessible pages, and used an AI-detection tool to assess the source text. They reported evidence of AI-generated sources in citations from all four systems. The paper describes this as evidence of AI-generated sources appearing in generative search citations—not proof that each flagged page was erroneous or spam.
As an Amazon Associate I earn from qualifying purchases.
About 16% of successfully scraped cited sources across all four engines were classified as AI-generated. That is a combined result, not a Perplexity-only rate, and it excludes cited pages the researchers could not successfully scrape. The figure therefore should not be read as the share of all Perplexity citations that are AI-generated. A secondary analysis also discusses the combined figure, but the underlying paper is the primary source.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteThe paper gives topic-specific figures for Perplexity: 237 sources, or 2.7% of citations, for health queries, and 60 sources, or 1.1%, for politics queries. These are results for those topics in that study, not a general estimate of Perplexity’s citations. The audit does not establish a platform-wide rate or a trend over time.
#1 Best Overall
Does AI-generated mean erroneous or spam?
No. “AI-generated,” “unsupported,” “erroneous,” and “spam” describe different things. The audit used an AI detector to classify source text; it did not establish that every classified page was false, low-quality, or created to manipulate search. A generated page can be accurate, while a human-written page can be wrong. The authors also describe limitations in AI-detection tools, so classification is evidence about likely provenance, not a definitive label for every page. The paper supports a narrower conclusion: source text classified as AI-generated appeared in the audited citations.
Can you trust Perplexity’s citations?
A citation is a lead to a source, not a guarantee that the source supports the exact claim beside it. Common Sense Media’s Youth AI Safety Institute makes that distinction in its 2026 Perplexity risk assessment: “A citation is not itself proof that Perplexity’s synthesized claim accurately represents the source.” In its assessment sample of 1,022 citations, the organization reported that 28% came from user-generated sites without editorial accountability and 28% from government agencies, universities, and peer-reviewed research. Those percentages describe that assessment’s sample; they are not Perplexity’s overall citation mix.
A separate audit checked citations for numerical claims
Haus Research tested Perplexity’s Sonar and Sonar Pro API models using 310 factual questions about 210 technology companies. In a September 2, 2026 report, it examined 1,826 citations attached to numerical claims. It found that 34.7% either did not open to an ordinary reader or opened without containing the number cited. The models were accessed through OpenRouter, and the result applies to that audit’s snapshot, subject area, and opening-or-number check—not to Perplexity’s consumer product or all of its citations. Haus Research also says this check does not establish whether the answer itself was true. Read the audit’s methods and qualifications.
Recommended Free Tools
This API audit and the AI-provenance study measure different things. One classified accessible source text for likely AI generation; the other tested whether citations attached to numerical claims opened and contained the cited number. Their percentages cannot be combined into a single estimate of unreliable Perplexity citations.
Rank #3
How to check a Perplexity citation
For a claim that matters, open the source and verify the specific statement rather than relying on the presence of a citation.
- Check the exact claim. Find the cited sentence, figure, or date on the linked page. A source about the same topic may still not support the answer’s wording or number.
- Check who published it and when. Look for a named organization or author, an editorial process, and a publication or update date relevant to the claim.
- Follow the evidence. For consequential factual claims, look for linked primary material such as a government document, university research, company filing, or original dataset.
- Compare independent sources. If the page does not contain the cited detail, is inaccessible, or relies on another unverified page, do not treat the citation as confirmation. Seek another source that directly supports the claim.
These checks address separate problems: a source’s likely authorship does not determine whether a claim is supported, and a citation that fails to support one number does not by itself prove the answer false.
Quick Recap
Best Value
Rank #4
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Free tools Windows power users keep installed
One-click scans. No signup required.




