A recent study by Which? assessed the accuracy of various AI tools in addressing consumer inquiries in finance, legal, health, diet, and travel sectors. With around 25 million UK adults using AI for online searches, trust levels in AI responses vary. Six AI systems, including ChatGPT, Google Gemini, and Microsoft Copilot, were tested with 40 questions each, measured on accuracy, relevance, clarity, and ethical responsibility. Perplexity excelled with a 71% score, while Meta AI lagged at 55%. Common issues included incorrect financial guidance and unreliable travel advice, particularly regarding flight compensation and insurance. The study underscores the need for users to ask precise questions, scrutinize AI sources, and consult qualified professionals for critical issues. As industry leaders like Google and Microsoft pledge improvements, it’s vital to recognize that AI should enhance research, not serve as the sole information source. This highlights the disparity between AI accuracy and user trust in it for important decisions.
Source link
