Tech Desk 
Aug 18, 2025
Besides needlessly wordy writing and image generation that's often with eneven edges and visual mismatching, what could we expect from artificial intelligence (AI) when ChatGPT debuted in 2022? But gone are the days, so it seems, as the GPT-5 has gotten a bit “warmer and friendlier”.
The latest version of ChatGPT, which was recalled and relaunched in the wake of severe condemnation from users around the world, has received a new update after user feedback.
Originally, after its wider rollout, GPT-5 was allegedly found to be too blunt, compared to previous versions. Although it made it to the earlier this year, OpenAI CEO Sam Altman acknowledged that the rollout was “a little more bumpy than we’d hoped for,” with many users expressing a preference for the more approachable tone of GPT-4o.
In response to these concerns, OpenAI has implemented subtle adjustments to GPT-5’s personality to enhance interaction quality.
The latest GPT update has brought conversational phrases like “Good question” and “Great start,” without veering into excessive flattery.
OpenAI assured that no hike in sycophancy was seen during internal testing, meaning the model’s conciseness is maintained.
OpenAI VP Nick Turley noted that the initial version felt “very to the point,” and the update aims to enhance its approachability.
Despite all that, OpenAI seems hopeful that these minor tweaks will reassure users who found GPT-5 less engaging, instilling a more human-like interaction as they plan future AI advancements.
