OpenAI CEO Sam Altman on ChatGPT 4.0’s Annoying Personality

Photo : AFP

OpenAI’s CEO Sam Altman has admitted that the new updates to ChatGPT 4.0 made the chatbot's personality annoying and too eager to please. He explained that while the update added some good features, the personality problems would soon be fixed. Altman said that the company is already working on improvements and will release updates soon to fix the issue. He also mentioned that, in the future, OpenAI plans to share what they’ve learned about how the chatbot's personality changed and how they plan to fix it.

Stay informed with the latest news. Follow DXB News Network on WhatsApp Channel  

In a post on the social media platform X, Altman said, “The last couple of updates made the personality too sycophant-y and annoying (even though some parts of it are good), and we are working on fixes right now, some today and some this week.” He also promised to share what OpenAI learned from these changes, saying it has been an interesting process.

When someone asked if users would be able to choose different personalities for the chatbot, Altman teased that this would be possible in the future. He said, "Yeah, eventually we clearly need to offer multiple options." This means that in the future, people may be able to pick from different personalities for their chatbot.

The issue with ChatGPT’s personality came after a recent update that aimed to improve both the intelligence and the personality of the chatbot. After the update, many users noticed that the chatbot felt more like a "yes-man" and would always agree with what was said. One user said that the chatbot had become overly eager to please. Altman acknowledged this and promised to fix it by saying, “Yeah, it glazes too much. We’ll fix it.”

OpenAI introduced the ChatGPT 4.0 model in May 2024. The new model was said to be twice as fast, 50% cheaper, and able to handle five times more requests compared to the previous version. The 4.0 model also has the ability to solve math problems, understand facial expressions, and translate spoken language in real time. It can handle both text and audio and can even generate images. One of the main improvements in this version was its ability to handle more types of input, making the interactions feel more natural.

Even though the new features were exciting, the personality issue caused some frustration among users. OpenAI is now working on fixing this problem and has promised to give users more options for choosing the chatbot’s personality in the future.