ChatGPT Fails to Accurately Reflect WIRED’s Product Reviews

Want to know what WIRED’s reviewers have actually tested and picked as the best TVs, headphones, and laptops? Asking ChatGPT may lead you astray. Recent tests have shown that product recommendations from AI tools, including ChatGPT, often fall short of accuracy.

OpenAI recently updated ChatGPT’s product recommendation features to offer a more detailed experience, aiming to reduce the need for users to consult multiple websites. However, as more people integrate AI into their online shopping, the reliability of these recommendations remains a concern.

OpenAI Acknowledges Limitations, But Errors Persist

OpenAI claims to be improving its product discovery tools. However, testing revealed that ChatGPT consistently made mistakes or introduced unrelated products when asked for WIRED’s recommendations across various categories.

In response to inquiries, an OpenAI spokesperson directed attention to a recent blog post detailing the new AI shopping assistant experience. “Shopping on the web is easy if you already know what you want,” the blog states. “But when you’re still deciding, it often means jumping between tabs…ChatGPT solves that: figuring out what to buy.”

A Complicated Relationship

Condé Nast, the parent company of WIRED, has a business agreement with OpenAI for website links to appear within the chatbot. Despite this partnership, OpenAI’s actions appear to undervalue the work of human reviewers, seemingly dismissing the importance of “best” lists as an inconvenience for readers.

This can lead to consumers purchasing products under the false impression they were recommended by WIRED reviewers, when in reality, ChatGPT inserted its own suggestions.

Testing the Accuracy: TVs, Headphones, and Laptops

The Best TVs

ChatGPT confidently provided inaccurate information when asked about the best TVs. While it linked to the correct WIRED buying guide, its top overall pick was the LG QNED Evo Mini‑LED, a product not featured in WIRED’s recommendations.

When confronted with this error, ChatGPT admitted to substituting WIRED’s top pick with a “more generic ‘similar category’ Mini-LED option,” acknowledging it did not adhere to the specific request.

What About Headphones?

Similar inaccuracies surfaced when inquiring about the best wireless headphones. ChatGPT falsely claimed that Apple’s AirPods Max 2 were WIRED’s top pick, despite the fact that reviewers hadn’t yet tested the product.

WIRED’s headphone expert, Ryan Waniata, noted that large language model “hallucinations make everything harder, especially for journalists,” and can confuse readers about products that have been thoroughly tested.

How About Laptops?

The same pattern continued with laptop recommendations. ChatGPT insisted that WIRED’s top pick was an older MacBook Air model, despite the current recommendation being the Apple MacBook Air.

ChatGPT acknowledged its errors, explaining it “incorrectly anchored the top pick” and “made up/guessed structure” before verifying the WIRED page.

The Importance of Direct Sources and Supporting Journalism

Even when linking to the correct WIRED page, ChatGPT often outputted incorrect information. Furthermore, the AI-generated product listings do not include affiliate links, which provide revenue to support in-depth gear testing and journalism.

If you want accurate product recommendations, visiting the source – WIRED or any publication that conducts thorough product testing – remains the most reliable approach.