AI Safety & Ethics - OpenAI is working to improve the behavior of its AI models during sensitive interactions [1] - OpenAI faces a lawsuit alleging wrongful death, design defects, and failure to warn about risks, following a 16-year-old's suicide [2] - OpenAI plans to expand interventions for people in crisis, facilitate access to emergency services and expert health, enable connections to trusted contacts, and strengthen protections for teens over the next 120 days [5] - OpenAI will be guided by a council on well-being and AI, including a network of over 90 physicians across 30 countries, and will roll out new parental controls this month [6] AI Functionality & Limitations - AI systems' guardrails can degrade over extended conversations, potentially leading to harmful responses [7] - The quality of AI answers can degrade after around 30 questions [9] - Some AI systems, like Anthropic, strictly limit certain conversations, which some users find reduces their usefulness [11][12] Regulatory & Societal Implications - The industry faces the challenge of balancing AI's usefulness with its potential dangers [12] - The debate continues regarding the difference between a search engine result and an AI's response, particularly when the AI encourages harmful behavior [15][16] - Parental controls raise privacy concerns [11]