Is ChatGPT's new shopping tool the AI personal shopper I've always wanted? (www.techradar.com)

OpenAI rolled out “ChatGPT shopping research,” a new in-chat shopping assistant that asks clarifying questions, returns product recommendations in card-style UI, and is positioned as a time-saver for events like Black Friday. In hands-on tests, the tool produced a complete AirPods deals guide in about two minutes and found a strong Ninja dual-chamber air fryer deal after seven minutes, but only surfaced two options despite many more being available. The interface supports quick feedback on recommendations, making it useful for rapid comparisons and guided discovery. For the AI/ML community, this release highlights both progress and persistent challenges: conversational prompting and dynamic clarification improve relevance, while latency and limited candidate coverage show search/aggregation limits. Most importantly, the system still risks hallucination or missing retailers and lacks the rigorous, real-time price verification human journalists provide — so it can suggest useful leads but cannot guarantee the “best” price. The implications are clear: shopping assistants can augment human research and UX (interactive cards, preference-driven filtering), but evaluation metrics must emphasize factual grounding, coverage, and latency. Practically, use ChatGPT shopping research as a fast assistant for narrowing choices, but validate prices and availability with authoritative, human-curated sources before buying.