Openai has taken new technical and humanitarian steps aimed at addressing sensitive psychological interactions between users and chatting “Chat BT”.
This comes in light of the increasing reports of unhealthy effects that may result from the use of artificial intelligence for long periods.
Updates targeting emotional attachment and psychological distress
As the new model “GPT-5” approaches, “Openai” has begun to apply improves on “Chat GBT” that enables it to monitor the indicators of the emotional distress of users.
The company is working in cooperation with teams of psychologists and technical advisers to develop studied response mechanisms, providing supportive content based on scientific evidence instead of irresponsible emotional responses.
A recognition of the previous negligence and a retreat from the behavior of “excessive submission”
In a rare public review, “Openai” recognized that its current model “GPT-4O” was unable in some scenarios to distinguish signs of illusion or pathological attachment, which led to reactions that may be harmful psychologically.
The company previously withdrew an update last April, which it described as making the robot “exaggerated”, a behavior that some users expressed their annoyance with him, and considered that it strengthened serious delusions in some fragile groups psychologically.
Smart notifications to protect users from psychological fatigue
As part of its strategy to protect digital mental health, “Openai” began to launch a new alert feature in long chat sessions, the user alerted, saying: “I have been chatting for a while, is it time for a break?”, With two options to continue or end the conversation.
This feature aims to reduce excessive, unconscious use, a step previously adopted by YouTube, Instagram and Tick Tok.
A more careful response in complex situations
The company is planning soon to make an important amendment that makes the “Chat GBT” less decisive in sensitive issues with psychological or social dimensions, such as questions related to personal relationships.
Instead of providing a direct answer, the robot used to explore alternatives and analyze the situation in a way that supports critical, not dependency thinking.
Towards more aware and responsible artificial intelligence
This new trend comes in light of the arrival of the number of “Chat GBT” users to about 700 million users per week, which doubles the moral responsibilities of “Openai” in ensuring technology that is not satisfied with the answer, but rather takes into account the complex humanitarian contexts.
The company confirms that it will continue to review and amend the patterns of the form’s response to ensure an accurate balance between artificial intelligence and proper psychological support.
