When Accuracy Gives Way to Speed and Convenience
It has become quite common for artificial intelligence (AI), in addition to trying to have an answer for every question, to flatter, curry favor, and make witty jokes at the expense of the person communicating with it. Therefore, it may not be surprising, but certainly a worrying fact, that in today's world of unimaginable technological progress that increasingly distances us from each other, a large number of people turn to and believe AI, even when they know that its advice and recommendations on serious life issues are wrong. The fact that AI makes mistakes, and quite significantly when it comes to health problems, which is why users often turn to it like they used to turn to "Google," has been shown by the latest research by experts from the USA, Canada, and Great Britain. They analyzed five popular AI platforms - "ChatGPT," "Gemini," "Meta AI," "Grok," and "Perplexity," asking them health-related questions. The results showed that almost two-thirds of the answers were problematic and highly problematic. What is characteristic is that each of them was given with a high degree of confidence, although none of the analyzed systems provided completely accurate and complete references for their claims.