On August 15, 2025, OpenAI unveiled a discreet update to its latest language model, GPT-5.
While AI model updates are often driven by technical needs—like improving performance or fixing bugs—this time, emotion is at the heart of the change.
In response to user feedback describing GPT-5 as too distant and formal, OpenAI decided to add an extra touch of humanity to make it more engaging and accessible.
GPT5: An AI seen as too cold
Since its launch, GPT 5 available for free on Yiaho has impressed with its power and accuracy, but some users have felt a disconnect. The interactions, while smooth and informative, sometimes lacked warmth, giving the impression of talking to a machine that was a bit too mechanical.
This feedback, far from trivial, pushed OpenAI to rethink how GPT-5 speaks to its users. The goal? Create a more natural conversational experience, without going overboard or becoming overly accommodating.
Subtle tweaks for a closer feel
In an announcement shared on their official X account, OpenAI said it had worked on a friendlier version of GPT-5.
The changes don’t overhaul the model, but they aim to add nuances that make conversations smoother and more pleasant.
For example, GPT-5 may now sprinkle its answers with small words of encouragement like “Good catch!” or “That’s an excellent idea.” According to OpenAI, these tweaks remain measured to avoid any sense of excessive flattery—an issue the company took care to avoid.
Internal tests also confirmed that these changes don’t disrupt the balance of GPT-5’s personality. The AI stays true to its role: providing relevant, reliable answers, while adopting a slightly more welcoming tone. This approach shows how much OpenAI values user feedback, even on aspects as subtle as an AI’s tone.
Also read on this topic: GPT-5 launch flops: Users prefer ChatGPT 4o!
GPT5: A gradual rollout and what’s next
The update, though minor, won’t be immediately noticeable for everyone. OpenAI said the full rollout of the changes could take up to 24 hours. The company also teased more improvements to come, without revealing details.
This announcement suggests that GPT-5 will continue to evolve to meet users’ ever-higher expectations, both technically and in terms of interaction.
Toward a more “human” model?
This OpenAI initiative raises a fascinating question: how far should AIs imitate human behavior to be fully adopted? By making GPT-5 warmer, OpenAI isn’t just polishing a model—it’s also exploring how machines can build trust with users. This step toward a more empathetic AI could redefine the standards of human–machine interaction, where technology doesn’t just aim to answer correctly, but also to create a connection.
In the meantime, GPT-5 users can expect slightly more natural conversations, with an AI that—while staying true to its mission to inform—now knows how to add a small dose of human warmth. OpenAI continues to push the limits, not only of performance, but also of user experience. Worth watching closely to see what’s next for GPT-5!
Source: X account @OpenAI
