WVU researcher says AI should disclose what it doesn’t know
Like humans, generative artificial intelligence doesn’t always know when it’s wrong, and a West Virginia University researcher is trying to get AI agents like ChatGPT to recognize — and admit — when that happens.