Kudos to OpenAI for listening to users and rolling back a recent change to its ChatGPT model. Users took to social media to complain about the AI’s defensive tone. They believed it had changed enough to change too severely, leading to the ruling. ChatGPT tone This year, most conversations about the new ChatGPT have focused on the tool’s tone. It highlights how important the user experience is to such interactions with AI.
The first of these updates to ChatGPT, which sought to add what was described as an “anti-sycophant-y, warmer and friendlier” demeanor. Not all users welcomed this change. Others lamented that the new model seemed much more sterile and less personal than iterations past. This feedback led OpenAI to reconsider some of the changes it had made to the AI’s tone and style.
In an era when using AI tools quickly becomes the default mode for how we all work, that balance is critical to strike. Striking this balance is critical to user satisfaction. These mixed results reflect the complicated and often unpredictable expectations users bring to their experience with an AI tool. Some users loved the conversational tone, while others found the casual voice to be off-putting and unprofessional.
OpenAI’s move to roll back the update is a testament both to the times and to OpenAI’s willingness to respond to its growing user base. By being transparent about these concerns, the company is trying to improve user experience and build trust in the ChatGPT model. The decision is a boost to an enduring effort to make AI behavior correspond with user intent.
“dark pattern” – techcrunch.com
