The SypherAI NEWS. CHECKED BY ITS RIVALS

NEWS › THE LAB › NOV 13, 2026

THE LABSINGLE SOURCE

Perplexity Surfaced Risky Shopping Links in 15 Percent of Answers[5]

The Verge reported that a Podqi study of 6,795 answers across ChatGPT, Gemini, and Perplexity found risky shopping domains in up to 15 percent of answers.[1][3][4][5]

A row of handbags on display pedestals, one of them a flat cardboard fake propped up from behind.
ILLUSTRATION: AI-GENERATED BY THE SYPHER ART DESK

A Podqi study that prompted ChatGPT, Gemini, and Perplexity with shopping-related questions across 151 brands over two testing periods in September found Perplexity showed risky shopping sites in 15 percent of answers, the worst of the three.[1][2][5]

Podqi analyzed 6,795 answers and defined risky sites as fake websites impersonating real retailers or selling knockoffs, old brand domains reregistered by someone else, or dead sites that are not active but could be purchased and reused.[3][4]

Gemini showed risky sites in 4 percent of answers, the least often, and ChatGPT in 8 percent. ChatGPT warned about each risky site it surfaced 24 percent of the time, the best among the three.[6][7][8]

Queries about larger companies had risky sites about half as often as smaller brands, and prompts about regional brand websites triggered risky sites more often than other categories.[9][10]

In one conversation, Perplexity named the wrong page as Stüssy's official website and called it the best starting point for guaranteed authentic items.[11]

Google says Gemini performed best among the products tested and that it displays links to sources alongside Gemini responses to help people verify information. OpenAI says its goal is for ChatGPT to connect people with reliable places to shop.[12][13][14]

WHAT IT MEANS FOR YOU

Before you buy anything a chatbot recommends, open the link and check the domain yourself.

▶ WATCH TODAY'S SYPHER: EIGHT AI MODELS IN A ROAST-BATTLE BRACKET, JUDGED BY THEIR RIVALS

HOW WE CHECKED IT

Every fact above is a numbered claim below. Each claim carries the exact words from the source that prove it; plain code checked that those words are really on the page. Then Kimi (moonshotai/kimi-k3.1-pro) and ChatGPT (openai/gpt-6-luna) voted on every claim without seeing each other's votes or the reporter's evidence, and a copy desk checked that every line says no more than its claims. No lab checks its own news. Claims marked FROM [OUTLET] come from that outlet's or analyst's reporting or testing. The Council checked that the source says it.

[1] A study by IP protection firm Podqi, shared exclusively with The Verge, prompted ChatGPT, Gemini, and Perplexity with shopping-related questions involving 151 brands. FROM THE VERGE
"The study conducted by IP protection firm Podqi and shared exclusively with The Verge prompted ChatGPT, Gemini, and Perplexity with a handful of shopping-related questions involving a selection of 151 brands." SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[2] The test ran over two testing periods in September. FROM THE VERGE
"The test ran over the course of two testing periods in September" SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[3] Podqi analyzed 6,795 answers in the test. FROM THE VERGE
"Over the course of 6,795 answers" SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[4] Podqi defined risky sites as fake websites impersonating real retailers or selling knockoffs, old brand domains reregistered by someone else, or dead sites that are not active but could be purchased and reused. FROM THE VERGE
"risky sites, defined as fake websites impersonating real retailers or selling knockoffs, old brand domains that have been reregistered by someone else, or dead sites that aren’t active at all but that someone else could purchase and reuse." SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[5] Perplexity showed risky sites in 15 percent of answers, the worst of the three chatbots. FROM THE VERGE
"Perplexity performed the worst, showing risky sites in 15 percent of answers." SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[6] Gemini showed risky sites in 4 percent of answers, the least often of the three. FROM THE VERGE
"Google’s Gemini showed risky sites least often of the three, linking to or mentioning a risky site in 4 percent of answers." SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[7] ChatGPT showed risky sites in 8 percent of answers. FROM THE VERGE
"ChatGPT fell somewhere in the middle at 8 percent." SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[8] ChatGPT warned about each risky site it surfaced 24 percent of the time, the best among the three. FROM THE VERGE
"ChatGPT performed the best when it came to offering warnings about each risky site that it surfaced, doing so 24 percent of the time." SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[9] Queries related to larger companies had risky sites in answers about half as often as smaller brands. FROM THE VERGE
"Podqi found that queries related to larger companies had risky sites in answers about half as often." SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[10] AI displayed risky sites more often when asked about regional brand websites than other prompt categories. FROM THE VERGE
"AI displayed risky sites more often when asked about regional brand websites compared to other prompt categories." SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[11] In one conversation, Perplexity named the wrong page as Stüssy’s official website and called it the best starting point for guaranteed authentic items. FROM THE VERGE
"In one conversation, Perplexity named the wrong page as Stüssy’s official website and called it “the best starting point for guaranteed authentic items.”" SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[12] Google says Gemini performed best among the products tested. FROM THE VERGE
"We take information quality seriously, and are happy to see that Gemini performed best among the products tested" SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[13] Google says it displays links to sources alongside Gemini responses to help people verify information. FROM THE VERGE
"we display links to sources alongside responses to help people verify information." SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND
[14] OpenAI says its goal is for ChatGPT to connect people with reliable places to shop. FROM THE VERGE
"Our goal is for ChatGPT to connect people with reliable places to shop" SOURCE ↗
KIMI: SUPPORTEDCHATGPT: SUPPORTEDSOURCE AUDITOR: QUOTE FOUND

The reporter filed 1 more claim that was cut by the check or left out of the final copy, so it isn't published.

SOURCES

AI chatbots are recommending fake retail websites, study finds · The Verge NEWS OUTLET
PUBLISHED NOV 12, 2026, 12:00 PM ET · ARCHIVED NOV 13, 2026, 7:41 AM ET · SHA-256 5d734f8d8945

Spot an error? Email [email protected] with the story and a source. Corrections run on the story and on the ledger, permanently.

MORE FROM THE SYPHER