OpenAI updates ChatGPT: Personalization, youth protections, and Codex improvements

OpenAI announced three notable updates for ChatGPT:

• Unified personalization & new personalities: Personalization controls have been consolidated into a single “Personalization” settings page. Users can choose or tweak different ChatGPT personalities to better match tone and behavior, and manage memory and preferences more easily.
• Enhanced youth protection: New safety measures restrict certain interactions for users under 18, including bans on flirtatious conversations and other safeguards to reduce exposure to inappropriate content.
• Codex / developer improvements: OpenAI introduced an optimized Codex variant (GPT-5-Codex) targeted at coding tasks and code reviews. It’s intended to be the default in cloud coding workflows and selectable for local Codex toolchains and IDE extensions. Some developers have reported regressions in certain coding tasks, so results may vary as the model is refined.

These changes aim to make ChatGPT more customizable and safer for younger users, while improving developer tooling—though the Codex update may still need refinement based on developer feedback.

Read the official release notes for full details: OpenAI ChatGPT release notes

What do you think—do the new personalities and protections improve ChatGPT for you, or do you have concerns? Leave a comment below.