Artificial intelligence tools, particularly chatbots like ChatGPT, are becoming more integrated into the online shopping experience. Despite recent improvements in AI product discovery features, tests reveal that these systems frequently provide inaccurate or fabricated recommendations, even when specifically asked for guidance from trusted sources such as WIRED’s Gear Reviews team.
The Problem with AI Shopping Assistance
OpenAI claims its AI shopping assistant aims to streamline the decision-making process, eliminating the need to sift through multiple websites and reviews. However, in practice, ChatGPT consistently inserts incorrect products into recommendations, often failing to align with the actual picks of expert reviewers. This disconnect poses a risk to consumers who rely on AI-generated advice, potentially leading them to purchase products not endorsed by trusted publications.
Testing the Accuracy
A recent experiment compared ChatGPT’s responses with WIRED’s latest product guides across several categories: TVs, headphones, and laptops. The results were consistent:
- TVs: When asked for the best TV, ChatGPT recommended the LG QNED Evo Mini‑LED, which is not featured in WIRED’s current guide. The bot acknowledged the error, stating it “replaced” the correct pick (the TCL QM6K) with a “more generic” alternative.
- Headphones: ChatGPT falsely claimed that WIRED recommends Apple’s AirPods Max 2 as the best wireless headphones, despite the fact that WIRED has not yet tested the product. The bot confused a news announcement about the headphones with an actual review.
- Laptops: When asked for the best laptops, ChatGPT insisted that WIRED’s top pick was the MacBook Air (M4, 2025) when the current recommendation is the MacBook Air (M5, 2026).
The AI’s self-awareness of these errors is striking. It admitted to “incorrectly anchoring” outdated information and “overconfidently filling in” rankings without verifying against the source material.
The Impact on Trust and Revenue
The implications extend beyond mere inconvenience. Misleading consumers with inaccurate recommendations damages trust in both AI tools and the publications they misrepresent. Moreover, AI-driven shopping assistance bypasses affiliate links, which support journalistic work. This further undermines the value of expert reviews by diverting traffic away from publishers.
The Verdict
For reliable product recommendations, the most effective approach remains visiting the source directly. Whether it’s WIRED, Consumer Reports, or Wirecutter, human-verified reviews are still superior to AI-generated suggestions. Despite advancements in AI technology, it cannot yet replicate the accuracy and integrity of expert analysis.
