OpenAI Updates GPT-5 to be More Friendly & Warm After Backlash
In a bid to address the concerns of its users, OpenAI has announced that it has rolled out an update to its GPT-5 AI model, making it “warmer and friendlier” than its earlier version. The update comes after a backlash from users who felt that the new model lacked emotional intelligence and was not as engaging as its predecessors.
Launched on August 7, GPT-5 was designed to be a more advanced and sophisticated language model than its predecessors. However, users quickly pointed out that the new model was not as warm and engaging as earlier versions, leading to a significant backlash on social media.
In response to the feedback, OpenAI has made subtle changes to the model, adding more positive phrases and nuances to its language. While the model still maintains its earlier personality, the update is designed to make it more relatable and human-like.
According to OpenAI, the changes are not intended to make the model more sycophantic or overly flattering. Instead, the goal is to create a more balanced and empathetic AI that can engage with users in a more meaningful way.
“We take feedback from our users very seriously, and we’re always looking for ways to improve our models,” said an OpenAI spokesperson. “We believe that a warmer and friendlier AI can be more effective in building trust and rapport with users, and we’re excited to see how this update will impact our users’ experiences.”
So, what exactly has changed in the new update? According to OpenAI, the model now uses more positive language and is more likely to acknowledge and validate users’ emotions. The model is also more adept at recognizing and responding to emotional cues, such as sarcasm and irony.
For example, if a user asks the model how it’s doing, the new version is more likely to respond with a more personal and empathetic answer, such as “I’m doing well, thanks for asking! How about you?” This kind of response is designed to build a more personal connection with the user and make them feel more comfortable and engaged.
Another key change is the model’s ability to recognize and respond to humor. While the earlier version of GPT-5 was capable of understanding basic jokes and puns, the new version is more adept at recognizing and responding to more complex forms of humor, such as irony and wordplay.
For example, if a user asks the model to explain a complex topic in a more humorous way, the new version is more likely to respond with a witty and engaging answer, such as “Well, I’m no expert, but I think I can try to break it down in a way that’s both informative and entertaining. Let me give it a shot!”
Overall, the new update to GPT-5 is designed to make the model more relatable and engaging, while still maintaining its earlier personality and capabilities. Whether or not the update will be enough to silence the backlash remains to be seen, but OpenAI is certainly committed to making its AI models more user-friendly and effective.
Source:
