. Chatgpt users A few days ago They noticed a very strange behavior in the use of openai chatbot. The result was the withdrawal of an update Even Sam Altman himselfOpenai CEO, He finally recognized the public openly. The update was applied directly to the GPT-4O model. Chatbot personality has changed seriously To such a degree to impress your answers. Something concerned that Chatgpt has already reached the search engine to a degree of wax.

Chatgpt begins to be strange in your answers, should we be afraid of it?

In particular, the personality of the GPT-4O model “Accepted and annoying” For users, Sam Altman accepted in a broadcast on X. Although users do not detect this, he stressed that he has “very good pieces”. He showed this They will result in “at some point” And “sharing” the learning obtained.

How did Chatgpt move with this update? Basically He answered in a “wrong” way with tonal rise Indeed, questions showing the response of feedback enough to produce others. In Spain only early in the morning, Chatgpt people were responsible for supporting this update.

The arrival of chatgpt as part of Apple Intelligence

A broader explanation of what is happening on Openai’s blog In this way, Openai performs comprehensive tests of its models. Before starting such an update. Adjustments made in the interior of the model They aimed to improve his personality in a way to “intuitive and effective”.. Based on their own model specifications, they are justified by showing these models to apply the principles with their own signals. “They focused on short -term comments too much without considering their evolution” As for user interactions over time.

An artificial intelligence model can withstand your personality more than your knowledge, where did we see it?

Chatgpt wanted to be more empathic with answers, but they didn’t have enough time for it.. Giving a personality to a chat boat requires education and exactly many evidence and errors before going to the public light and causing something as seen before. It is enough to remember what happened in Avengers in Fiction: Ultron Era, how A simple creation for good can do evil.

People Openai shows that you want to continue betting on personality options despite User will lead to feedback Creating new personalities. They will continue to look for this sincere and comprehensive personality, but this has proven that nothing can be shed light on the updates.

CATGPT is available in more than 40 countries

Openai established an action plan 4 The place they committed to continue working at the basic point: Education and system indicators, Create protection barriersexpand Users may have tests and this feedback Before and Expand the assessments according to their own characteristics and research To avoid future personality problems in models. Can they achieve a more aggressive and flattened personality in the future?

You can follow iPadizat Whatsapp on Facebook, Twitter (x) Or consult our Telegram channel to be up -to -date with the latest technology news.