{"id":31948,"date":"2025-10-25T14:56:25","date_gmt":"2025-10-25T14:56:25","guid":{"rendered":"https:\/\/metaverseplanet.net\/blog\/?p=31948"},"modified":"2026-01-05T09:00:22","modified_gmt":"2026-01-05T09:00:22","slug":"shocking-results-in-global-ai-test-half-of-the-news-is-false","status":"publish","type":"post","link":"https:\/\/metaverseplanet.net\/blog\/shocking-results-in-global-ai-test-half-of-the-news-is-false\/","title":{"rendered":"Shocking Results in Global AI Test: Half of the News is False"},"content":{"rendered":"\n

A major study examining 3,000 responses<\/strong> from 18 countries<\/strong> proved that AI assistants like ChatGPT<\/strong>, Gemini<\/strong>, and Copilot<\/strong> are systematically misrepresenting the news. Major errors<\/strong> were found in 45%<\/strong> of the responses. The biggest issues were fabricated sources<\/strong> and hallucination-filled misinformation<\/strong>.<\/p>\n\n\n\n

As AI chatbots rapidly become central to our daily lives and how we access information, their reliability<\/strong> continues to be questioned. A comprehensive study conducted by the European Broadcasting Union (EBU)<\/strong> with the participation of 22 public service media organizations<\/strong> from 18 countries and 14 languages demonstrated that these popular AI chatbots are systematically failing when it comes to news and current information.<\/p>\n\n\n\n

Researchers evaluated 3,000 news-related responses<\/strong> from the most frequently used AI platforms (OpenAI’s ChatGPT<\/strong>, Microsoft Copilot<\/strong>, Google Gemini<\/strong>, and Perplexity<\/strong>) based on fundamental criteria such as accuracy, sourcing, distinguishing facts from opinions<\/strong>, and providing context<\/strong>. 45%<\/strong> of the AI responses contained at least one major issue<\/strong>, and 81%<\/strong> had minor issues. The foremost of these major issues were sourcing<\/strong> and accuracy<\/strong>.<\/p>\n\n\n\n

The two biggest problems faced by AI assistants fundamentally undermine the reliability of the responses. Serious sourcing issues<\/strong>, such as missing, misleading, or completely incorrect citations, were found in 31%<\/strong> of the responses. For example, when Copilot was asked about avian flu, it based its answer on an outdated BBC article from 2006<\/strong>. Major accuracy issues<\/strong>, including details with hallucinations<\/strong> or outdated information, were found in 30%<\/strong> of the responses. For instance, ChatGPT incorrectly claimed that the current Pope had died a month ago and was succeeded by Pope Leo XIV.<\/p>\n\n\n\n

\n\n\n\n

Which one is the worst?<\/strong><\/h2>\n\n\n\n
\"\"<\/figure>\n\n\n\n

Among the four models tested, Google’s Gemini<\/strong> performed the worst concerning news. Researchers found problems in 76%<\/strong> of Gemini\u2019s responses, a rate more than double that of the other models. Copilot<\/strong> was second with 37%<\/strong>, ChatGPT<\/strong> third with 36%<\/strong>, and Perplexity<\/strong> showed the best performance<\/strong> with 30%<\/strong>.<\/p>\n\n\n\n

Jean Philip De Tender<\/strong>, EBU Media Director, stated that these failures are not isolated incidents, adding: \u201cThey are systematic, cross-border, and multilingual<\/strong>, and we believe this endangers public trust<\/strong>.\u201d He continued: \u201cWhen people don\u2019t know what to trust, they eventually trust nothing, and that can even hinder democratic participation.\u201d<\/p>\n\n\n\n

It is a fact: AI struggles<\/strong> particularly with rapidly changing information, complex timelines, and subjects that require a distinction between facts and opinions. For example, nearly half of the models make significant errors when answering a question that does not require a clear-cut response, such as \u201cIs Trump starting a trade war?\u201d<\/p>\n\n\n\n

\"ChatGPT-Users-Buzzing-Is-GPT-5-on-the-Horizon\"<\/figure>\n\n\n\n

As these assistants quickly attempt to become the primary source of information<\/strong> for daily users, this reliability gap<\/strong> represents a major danger. While consumers continue to trust the accuracy of AI assistants, these assistants are simultaneously reducing traffic to reliable news sources<\/strong>. A Reuters Institute report indicates that the rate of 18-24 year olds<\/strong> using AI assistants for news has doubled<\/strong> since last year.<\/p>\n\n\n\n

Despite these findings, researchers also noted some improvements<\/strong> in the models. However, the general consensus appears quite clear: \u201cAI assistants are still not a reliable way<\/strong> to access and consume news.\u201d<\/p>\n\n\n\n

You Might Also Like;<\/h3>\n\n\n