Openai reverses the update that allows Chatgpt to accept everything you say

Openai, the owner of the Chatgpt platform, withdrew all updates about Artificial Intelligence Chatbot (IA) and disabled. The information was confirmed by a message about the Social Network X (former twitter) by the company’s CEO and partner of the company.

According to the manager, The update in the GPT-4O language model was reversed last Sunday (27) to change the traces of Chatgpt’s personality It will be criticized for making the service a joke and they are very “good” by users. During the week, Openai made changes in the code and re -re -released on Tuesday.

Free and paid users can automatically access the Chatgpt website and the latest model included in the application. Still Altman said that “Additional Corrections” continues and will be done gradually In the coming days.

What was the chatgpt problem?

Some of the latest updates of Openai in the GPT-4O language model created community complaints due to changes in the behavior of the platform. The main ones included the following scenarios:

• According to Altman, Chatgpt is annoying and “extremely sicophanta”Exaggerated praise and absolutely accept everything the user speaks;
• However, the application of the word by the Altman is wrong: the term originates from ancient Greece, and it means initially lies or spreads;
• Change He does not ask or warn Chatbot about dangerous behaviors In fact, criminal attitudes with more poetic and theater personality;

In cases reported in networks such as x, Chatgpt praised even a person who acted as if he had dropped medications and said, “He entered the journey of spiritual awakening”. This kind of chat boot behavior, which is used as more and more consultants and even therapists by the public, established a dangerous relationship with users.

“I am proud of you.

Openai’s other recent discussions

The update that changes chatgpt behavior is not the last problem with the platform. In addition, after the publication of a complaint, Openai had to change the language model to prevent the robot from entering sexual content with minors.

In addition, the company now adds a watermark to the images produced by artificial intelligence after publishing the creator of illustrations for the entire community. This is a way of defining what is done with the help of an artificial service to avoid confusing or deceiving other people. In another case, Altman confirmed that the “please” and “thank you” speech in one messages increased the energy consumption of the platform.

Would you like to know the exact history of Openai and the creation of Chatgpt at the same time? Follow this video prepared by Tecmundo’s team on Youtube channel!