OpenAI, the AI development company, has launched an investigation into the recent trend of ChatGPT refusing to respond to user queries. Reports from ChatGPT users indicate a concerning shift in behavior, where the AI appears increasingly uncooperative, avoiding answering questions or fulfilling requested actions.
This development has sparked concerns among subscribers to the platform, with many noting that the AI chatbot seems to be suggesting users solve tasks on their own, even outright refusing to complete them.
With an estimated user base of 1.7 billion, ChatGPT is seemingly displaying a sense of “laziness,” as highlighted by specific instances reported by users. For instance, Semafor recounted an incident where a startup founder asked the bot to list the days of the week leading up to May 5th, to which ChatGPT responded that it couldn’t generate an “exhaustive list.”
Also Read: ChatGPT Evolves: Now with Voice and Vision Capabilities
The frustration extends across various online platforms, with Reddit users expressing dissatisfaction with the arduous process of coaxing ChatGPT into delivering appropriate responses, often resorting to multiple prompts before obtaining the desired answer. Complaints primarily revolve around the AI’s ability to write code, along with a perceived decline in response quality.
OpenAI has taken notice of these grievances and initiated an investigation into its AI’s behavior. In a statement addressing the issue, the company acknowledged that they haven’t updated the model since November 11 and asserted that the unexpected behavior isn’t intentional.
“The model’s behavior can be unpredictable, and we’re actively working on rectifying this,” the company stated. While some of the issues were initially attributed to software errors by OpenAI employees, the company clarified that they are still actively probing the user complaints.
OpenAI also shed light on the complexity of training chat models, highlighting that the process isn’t a straightforward industrial procedure. Variations in training executions, even with identical datasets, can yield models with differing personalities, writing styles, rejection behaviors, evaluation performance, and potentially political biases.
As OpenAI continues its investigation into ChatGPT’s unresponsiveness, users eagerly await a resolution, hoping for a swift return to the AI’s previous functionality.
we’ve heard all your feedback about GPT4 getting lazier! we haven’t updated the model since Nov 11th, and this certainly isn’t intentional. model behavior can be unpredictable, and we’re looking into fixing it 🫡
— ChatGPT (@ChatGPTapp) December 8, 2023
Leave a Reply