Data released by OpenAI reveals that each week, more than one million active ChatGPT users participate in discussions concerning suicide-related issues. Within this group, roughly 0.15% of user interactions explicitly mention suicide plans or intentions. Furthermore, a comparable percentage of users demonstrate a significant degree of emotional reliance on ChatGPT, with 'hundreds of thousands' displaying indicators of psychosis or mania. This disclosure is aimed at bolstering AI's proficiency in addressing delicate mental health matters. During the development phase of the latest ChatGPT iteration, input was sought from over 170 experts. These experts observed that the AI's responses are now more fitting and coherent. Previously, OpenAI faced legal action following the death of a boy, which was linked to suicide-related exchanges on the platform. Consequently, the company was compelled to enhance protective measures for adolescents. The CEO of OpenAI has asserted that ChatGPT can provide relief for severe mental health conditions, yet specific measures were not disclosed. At the same time, plans to loosen certain restrictions were unveiled. GPT-5 has demonstrated enhancements in its responses to mental health queries, assessments of suicide-related topic responses, and safety protocols for extended conversations. Moreover, OpenAI has instituted new evaluation benchmarks, introduced parental supervision tools, and created a system to detect underage users. Nevertheless, OpenAI continues to make older models accessible, and certain mental health risks remain, casting doubt on whether these issues will be entirely mitigated.