AI hallucinations are getting worse – and they're here to stay

AI chatbots just got a reasoning upgrade—but instead of getting smarter, they’re hallucinating more. OpenAI’s new o3 and o4-mini models, released in April 2025, are making up facts nearly 2 to 3 times more often than their predecessor. That’s a big red flag for industries like paper packaging, where AI is increasingly used to analyze regulations, sustainability data, and supply chain trends. If your chatbot misreads an FSC certification rule or fabricates recycling stats, that’s not just a glitch—it’s a compliance risk. The takeaway? AI’s not your oracle yet—fact-check everything.https://www.newscientist.com/article/2479545-ai-hallucinations-are-getting-worse-and-theyre-here-to-stay/

Comments

Popular posts from this blog

Coffee pods, bulky rigids may lose 'check locally' label

New DOJ proposal still calls for Google to divest Chrome, but allows for AI investments