OpenAI has announced that in the coming weeks it will tighten ChatGPT's restrictions on violent content and introduce additional safeguards for minor users, particularly concerning weapons and security system information. Under the new measures, when a conversation indicates that a user is persistently focused on violence or seriously harming others, ChatGPT may terminate the conversation on its own. Users whose conversations are terminated for this reason on multiple occasions may face account bans.