Users are also complaining about the quality of responses and the bot's ability to write code. Some want the company to return to the original GPT models. OpenAI has attributed some of the issues to a software bug and is still investigating user complaints. The company has stressed that the training process can result in different personalities for its models, including differences in writing style, refusal behavior, and even political bias.
Key takeaways:
- OpenAI is investigating reports that its AI bot, ChatGPT, is becoming more difficult to work with and is refusing to complete tasks.
- ChatGPT, which has 1.7 billion users, has been praised as a tool for improving efficiency, but recent user feedback suggests the bot's behavior is becoming unpredictable.
- Users have reported issues with the bot's ability to write code and have expressed a desire for the company to return to the original GPT models.
- OpenAI has stated that the training process can result in different personalities for its models, and that it is still investigating user complaints.