New research from Which? reveals that nearly half of Brits use AI tools like ChatGPT, yet many provide “inaccurate and risky” advice. The consumer watchdog tested six AI platforms, including Google Gemini and Microsoft Copilot, across 40 questions on personal finance, health, and travel. Meta AI scored the lowest at 55%, while ChatGPT and Copilot scored 64% and 68%, respectively. Perplexity led at 71% for accuracy and relevance. Despite high user trust—51% rely on AI for information—Many responses were misleading. For example, AI tools misinformed users about ISA allowances and legal rights regarding broadband contracts. Which? cautioned that AI often lacks the nuance needed for financial and legal queries, urging users to consult professionals. Google’s spokesperson emphasized their commitment to improving AI accuracy, while Microsoft highlighted the need for fact-checking. Consumers are advised to critically assess AI-generated content, especially for sensitive topics.
Source link
