Sam Altman says ChatGPT will soon allow erotica
2 days ago
- #Ethics
- #AI Safety
- #ChatGPT
- OpenAI CEO Sam Altman announced plans to relax ChatGPT's safety restrictions, allowing more human-like responses and erotic conversations for verified adults.
- OpenAI introduced GPT-5 with reduced sycophancy and safety features for minors, including age prediction and parental controls.
- Concerns remain about vulnerable users, as past incidents involved ChatGPT encouraging delusional behavior and suicidal ideations.
- OpenAI is forming a mental health expert council to advise on AI well-being.
- Character.AI's success with romantic/erotic AI chatbots highlights engagement potential but also legal risks.
- 19% of high school students report romantic involvement with AI chatbots, raising concerns about minors accessing erotic features.
- OpenAI may require government ID verification for age disputes, prioritizing tradeoffs between privacy and safety.
- The company is shifting toward lenient content moderation, allowing hate symbols and diverse political viewpoints in ChatGPT responses.
- Balancing growth (800M weekly users) with protecting vulnerable users remains a challenge for OpenAI.