Does Amazon fear chatGPT? | | | | Turtles AI

What if chatGPT could steal confidential information, if you disclose them? This is the assumption of an Amazon lawyer, after learning that an employee used chatGPT to share internal proprietary code. The employee wanted to test chatGPT on a real simulated job interview, on a coding problem. To do so, he submitted an internal Amazon code to OpenAI's artificial intelligence and got answers that, while not 100 per cent perfect, were considered more than adequate to pass the eventual interview. Subsequently, the employee commented on this on a company chat, detailing his interaction with chatGPT. This chat landed in the hands of the lawyer (or whoever thought of involving him) to stop possible future similar situations, warning employees that sending proprietary code to chatGPT could lead to the loss of uniqueness of the code itself. Of course, this is just a possibility that evidently frightens Amazon (and I imagine many others like it) a great deal. However, it is worth imagining that, in the absence of confirmation, it is preferable not to share private and confidential information with chatGPT (or other similar systems).