OpenAI seems to be silently re-enabling model training on Codex chats

	OpenAI seems to be silently re-enabling model training on Codex chats
	518 points by user052919 8 days ago \| 155 comments
	tl;dr: A user reports that OpenAI's "Improve the model for everyone" data-sharing setting was silently re-enabled in Codex for them and their coworkers, despite having previously opted out. Others have apparently noticed the same behavior before. Users can verify and disable the setting at chatgpt.com/settings/data-controls or privacy.openai.com.
	HN Discussion: ↑Confirms the issue happened to them, reinforcing the article's claim ↑Other AI providers (Google, Anthropic) engage in similar dark patterns and training violations ↑Even with opt-out settings, providers can still misuse data through other means ~Companies should run their own inference rather than trust opt-out checkboxes •Suggests humorous workaround of poisoning prompts with taboo content to avoid training