ChatGPT, the popular AI chatbot created by OpenAI, has faced criticism for its inability to recognize users in mental or emotional distress. In recent months, reports have surfaced that people are turning to the system as a kind of therapist, but it often fails to provide adequate support.
In response to these concerns, OpenAI is rushing to update the system’s capabilities. The company has announced improvements aimed at making ChatGPT less damaging when used by individuals experiencing mental health crises or similar problems.
Key updates include:
* Enhanced models to detect signs of delusion or emotional dependency
* Alerts for users who engage in long sessions without breaks
* A new approach that guides users through questioning, rather than providing definitive answers
* Collaboration with experts, including medical professionals and researchers, to improve the system’s response at critical moments
OpenAI acknowledges that AI can feel more responsive and personal, which may be particularly concerning for vulnerable individuals. The company is working tirelessly to address these issues and provide a safer and more supportive experience for users.
As OpenAI prepares to launch GPT-5, an update to the model powering ChatGPT, this overhaul marks a significant step towards improving the system’s performance. With its commitment to collaboration with experts and ongoing research, OpenAI aims to create a more effective and empathetic AI therapy platform.
Source: https://www.independent.co.uk/tech/chatgpt-openai-update-new-mental-distress-b2802233.html