The amount of time spent with generative AI helpers is rising. Actually, there are currently close to 700 million active users on OpenAI's ChatGPT each week. OpenAI stressed in a blog post today that ChatGPT was not created to hold users' attention indefinitely but rather to assist them in moving forward, learning something new, or resolving issues.
ChatGPT adopts a distinct strategy, concentrating on whether customers depart after achieving their objective, in contrast to the majority of online businesses that gauge performance using measures like time spent or clicks. In order to gauge how helpful customers find the service, OpenAI also tracks how often they return—daily, weekly, or monthly. Today, OpenAI revealed the following modifications to ChatGPT in order to better meet these objectives:
Better Distress Detection: Although ChatGPT is taught to be truthful, the GPT-4o model has occasionally failed to identify symptoms of emotional dependence or hallucination. In order for ChatGPT to react more suitably, OpenAI is enhancing the model's capacity to identify indications of mental or emotional distress.
Break Reminders: To promote taking breaks during extended user sessions, ChatGPT will now gently remind users to take breaks. To make these reminders seem helpful and natural, OpenAI will keep improving their appearance.
Managing Personal Inquiries: "Should I break up with my boyfriend?" is one of the many extremely personal inquiries that people ask ChatGPT. ChatGPT will refrain from providing conclusive responses in certain situations. Instead, by posing queries and balancing the advantages and disadvantages, it will assist consumers in making thoughtful judgments. This new approach to managing important personal choices will shortly be implemented.
OpenAI worked with subject-matter experts to create these behaviors in order to enhance ChatGPT's reaction in delicate circumstances. To enhance answers to questions about medicine, for instance, the team collaborated with over 90 doctors from more than 30 nations. Additionally, OpenAI stressed that this is a continuous endeavor and that it will provide additional information as it moves forward.