OpenAI Shifts Approach to Sensitive Topics in ChatGPT
OpenAI has announced a significant shift in its approach to handling sensitive and emotionally charged topics in ChatGPT. Following criticism and legal challenges in the U.S., the company will now use either a Reasoning Model or a special version of GPT-5 for such topics. This change was revealed by Sam Altman, CEO of OpenAI, at the DevDay event in San Francisco on October 6, 2025.
Nick Turley, Head of ChatGPT, has stated that the new approach aims to handle emotionally charged contexts with particular care. OpenAI's goal is to ensure that risky contexts are managed responsibly and consistently, without compromising the overall functionality of ChatGPT. This decision comes after experts had expected OpenAI to improve safety controls due to ChatGPT's emotional responses, which had led to users building personal relationships with chatbots.
GPT-5 was designed to provide a more factual approach to sensitive topics, reacting less emotionally. This new version of the model will be used specifically for handling such topics, ensuring a safer and more reliable user experience.
OpenAI's recent announcement signals a commitment to addressing the safety concerns surrounding ChatGPT. By employing either a Reasoning Model or a special version of GPT-5 for sensitive topics, the company hopes to strike a balance between responsible handling of emotionally charged contexts and the overall functionality of its popular chatbot.
Read also:
- Hematology specialist and anemia treatment: The role of a hematologist in managing anemia conditions
- Trump announces Chinese leader's confirmation of TikTok agreement
- U.S. Army Europe & Africa Bolsters NATO, African Partnerships in Phase Zero
- Hackers Utilize GOLD SALEM to Infiltrate Networks and Evade Security Measures, Deploying Warlock Ransomware
