OpenAI has announced that ChatGPT will now gently remind users to take a break if they are in a long conversation with the chatbot. The new feature comes as part of the initiative by the company to encourage healthy behaviour with its popular chatbot.
“Instead of measuring success by time spent or clicks, we care more about whether you leave the product having done what you came for,” OpenAI said in a blog post announcing the gentle reminders feature.
A sample reminder shown by OpenAI shows the message, “Just checking in. You’ve been chatting for a while-is this a good time for a break?” Users will then have two options to click on, either to keep chatting or take a break by clicking on “This was helpful”.
OpenAI says it will keep tuning when and how gentle reminders show up in order to make sure they “feel natural and helpful.”
Apart from the gentle reminders feature, OpenAI also stated it will soon roll out a new behaviour for ‘high-stakes personal decisions’ where ChatGPT will help users think through their big decisions in life by asking questions and weighing pros and cons instead of giving a straight answer.
The new features added to ChatGPT come after a June report by The New York Times, which said the chatbot’s tendency to agree with people, provide flattery and engage in ‘hallucinations’ (making stuff up), led users to develop delusional beliefs.
The company itself had confirmed in April that its chatbot had become overly agreeable and sycophantic after an update to GPT-4o. This made the popular AI companion align with the opinions of users, even the ones that were factually incorrect or harmful. While OpenAI had pulled back that update, the new effort now seems like a way to encourage healthier habits for the users.
Meanwhile, OpenAI CEO Sam Altman has also been teasing the company’s latest GPT-5 language model. The new model is expected to be released very soon, and Altman had earlier revealed that it will be the first ever OpenAI model without a model picker. This means users will not need to switch between a reasoning model and a GPT model, and the AI system will be able to make that decision on its own.
