In a new national study, Product AI evaluated four popular artificial‑intelligence assistants—ChatGPT, Gemini, Claude and Perplexity—using 220 shopping‑related questions across nine product categories. About 40% of consumers now rely on AI for shopping, according to the firm.

Methodology and key findings

Researchers asked each platform questions ranging from pricing details to product comparisons. The study uncovered numerous errors, including MacBook prices off by $300, sunscreen SPF ratings misreported by ten points, and earbud battery life estimates wrong by six hours.

Despite the mistakes, performance varied. Perplexity emerged as the top performer, with the company noting its ability to actively search the internet and other AI models before responding. Gemini ranked last, with more than half of its answers containing costly errors for shoppers, affecting both free and paid tiers. Claude placed third, while ChatGPT secured second place.

Industry reactions

Google, the owner of Gemini, disputed the relevance of the study, stating that Product AI tested a developer‑focused version of Gemini rather than the consumer app that powers everyday shopping experiences. The company emphasized that the Gemini app integrates billions of constantly refreshed product listings.

Amazon’s own AI shopping assistant, Rufus, continues to provide product and pricing data on the retailer’s platform, while Amazon has blocked other AI bots from searching its marketplace. TikTok has also entered the arena with a new AI shopping tool.

Consumer advice

Product AI advises shoppers to double‑check product information and pricing, especially when using AI tools that may still produce errors. As AI models evolve, accuracy can improve, but verification remains essential for protecting consumers from costly mistakes.

Original reporting: Allentown News – 6abc Philadelphia — read the source article.