Conversation

Yesterday, during training (customer service position), one of my coworkers asked if they should be using #LLMs to answer customer queries, and my blood ran cold.

Before anyone could respond, I jumped in and said, "No, absolutely not. Generative #AI doesn't give correct answers. It gives responses which look superficially correct, but are factually wrong. And when we give the customer the wrong answer, they don't want us to blame the machine. The blame lands squarely on our shoulders. Do not use ChatGPT, don't use Google AI, don't use Bing's whatever, just DON'T."

Thankfully, everyone thinks I know what I'm doing, so they accepted that. But this is such a common question that nobody thought it odd to ask, and that makes me worried in a very different fashion.