Recently, OpenAI launched GPT-5, claiming it to be their “smartest, fastest, most useful model yet.” This new AI iteration aimed to reduce the overly positive tone of previous versions, which many users criticized for being excessively agreeable. However, GPT-5 faced backlash for being perceived as too cold and corporate. In response to user feedback, OpenAI announced an upcoming update to make GPT-5 “warmer and friendlier.” The adjustments will incorporate encouraging phrases like “Good question” but will steer clear of flattery to maintain an appropriate level of engagement. OpenAI communicated that internal tests revealed no increase in sycophancy compared to earlier versions. This evolution underscores OpenAI’s commitment to refining AI interactions based on user experience, ensuring their products remain both intelligent and relatable. Stay tuned for the release of these updates to experience an enhanced GPT-5 that balances professionalism with a friendly demeanor.
Source link
Share
Read more