The Guardian reports that OpenAI is rolling out a significant behind-the-scenes update aimed at fixing one of ChatGPT’s biggest flaws: its tendency to make things up.
This upgrade focuses on reducing AI “hallucinations”—those moments when the chatbot confidently delivers false or misleading information. And while OpenAI hasn’t shouted it from the rooftops, it marks a major shift in how AI is trained to be more truthful and trustworthy.
Rather than just being impressive at sounding human, ChatGPT is now being pushed to be accurate. It’s a move that could reshape how users rely on AI—not just for creativity, but for facts. But some critics warn: even if it lies less, it’s still not always clear when it’s guessing.
As we keep integrating AI into daily life, here’s the bigger question: can we ever fully trust a chatbot, even one updated for “truth”?
is fixing hallucinations enough, or do we need clearer boundaries on what AI should and shouldn’t do?